How Do You Determine Class?

[traynor]

I have a major problem buying into the correlations of "class" and "speed." It is easy, convenient, and doesn't require much thought or effort to average a bunch of races and declare that because "better horses run faster" one can easily categorize the "class" of a horse by fitting it into a pigeonhole on a par chart. It may have made Quirin's algorithms to generate "par times" for various "classes" of races simpler, or Beyer's "projected speed" figures easier to handle, but I am highly skeptical that the massaging of various speed and pace numbers generates anything useful in the area of "class."

I have no comment to make on BRIS pars (or anyone else's) because I don't use them. I tried everything available some years ago, decided the logic used in creating them was seriously deficient, and the only solution (for me, and for my own wagering) was to create my own class level charts based on "other criteria."

A primary consideration in that endeavor was(is) an in-depth analysis of race conditions at a given track, in particular the ability of a horse in good form to perform competitively in "adjacent" class levels. Speed and pace figures were completely ignored in the determinations. I don't do stuff like this to keep an idle mind busy. I do it because it is profitable.

So the question is, what process do you personally use to determine if a given horse is going up in class, down in class, or is entered at essentially the same class level as in the immediately preceding races? And--once that determination has been made (if in fact it is made), what do you do with it?

[Overlay]

I find that purse or earnings formulas are still useful in class determination, particularly if they also take consistency into account, and are supplemented in the proper proportion by other metrics addressing speed, running style, and condition.

[Augenj]

Race purse, current and past.

[Robert Fischer]

Classy horses are talented, and they have the qualities that are required to run the race on their own terms.

Talent is speed. Running the race on their own terms allows them to fully express their talent.

Class will negatively affect a "cheap" horse by way of adversity.

[BlueChip@DRF]

Average earnings per start would be a good place to start.

[CincyHorseplayer]

The first thing I look at with any race is conditions and class ability of all the entrants with wins and up close finishes and as a thing in itself without any reference to speed or pace figures(at least initially).The location,surface,level,the movement around,it IS how you get to know the personality of the horse and the trainer.

What to do with it.I love outrun lines.It makes a horse shipping in from Hoosier Park to Beulah win at 4-1 or 10-1 because outrun speed competitive in today's race,instead of being in the vacuum of a faster race the horse finds itself in the ideal spot and gets aggressive.I love people who boast of never looking at past performances.It makes a horse that nearly won on OC62/N2X that is competing in an OC32/N1X win at 6-1.It makes layoff horse that nearly won a G3 race as a 2 year old 6 months ago win a N1X at 4-1 because it's speed figure is 6-8 points below the top in the field.Class is always relative to the field,but when comparing competitive speed the horse that earns it vs better is nearly always better than those who earn it vs lesser.I think only a few of us believe in it because it is such a race specific concept.It is only vaguely testable in a database.The latter being the criteria for belief,therefore it is not believed,therefore it is an overlay.

I'm not sure if I completely grasp where you were going with your post but this is the best answer I can give right now.Things like this arouse my interest because it fits neatly into my theory of all races as single factor events.1 factor that dominates a specific race and sits atop the hierarchy and renders secondary all other factors.And it can be so dominant that it can render 4-5 positive factors meaningless and underplayed.A single factor can shatter any theory of what is positive in any handicapping theory,in any database,single race dependant.Class is definitely one of those factors.And the more it is not believed the bigger it will indeed be overlayed.I like to look at it in isolation like I think you are saying you do.

[MJC922]

Having wrestled for many many years with time as the measure of class I eventually had to walk away from it. It was a difficult choice because I believe in a perfect world the clock will in fact allow us to capture class within the effort (as pace) however IMO there are currently too many other factors around to obscure this and it's not likely to change anytime soon (if ever), factors like wind, humidity, surface speed and then the real kicker are incorrect timings which my last research in this area flagged literally thousands per year on an annual basis (that is if you're breaking it down to interior fractions). So I mean absolutely no disrespect to those who stay on in the speed figure battles, but the best thing I did for myself was to get away from time. These days I generate class ratings instead of speed, horse against horse stuff, not Jim Quinn's class of the field ratings but ones of my own design. If we can rate an effort for class that's a start, I'm sure there are many ways other than mine. Though I haven't researched which efforts to use I have for years leaned toward using two relatively recent efforts. My leanings these days are to use class as an early elimination, and then move more deeply into current form to separate contenders, bringing in supplemental info from replays and physical appearance, and last but not least the odds I'm being offered will quite often be the deciding factor in which one gets the final investment especially in contentious races.

[Robert Goren]

There is no easy way. You have to do your own home work and compare how horses run when they change claiming prices and conditions. I was going to do some research into, but this damned dizziness put a stop to that.:bang: I still think there might be a market for class numbers similar to Beyers for speed, but it just not going to be me who develops it.

[PhantomOnTour]

First - how does one define class?

>the ability to carry ones speed thru fast splits and battle on?
>a horse who responds when challenged (even if he eventually loses)and takes up the bit...ala Shackleford or Silver Charm ?
>a horse who has won or been competitive vs higher levels than today's opposition?
>a horse merely coming out of a fast race?
>a horse merely coming out of a race at a higher level?
>bloodlines?

My opinion of class may differ from many, but to me class is the fight a horse puts up even if he's tired or in against supposedly classier rivals.
Class is the horse who always responds when they come to challenge him, no matter the surface, distance, or competition.

[traynor]

There is no easy way. You have to do your own home work and compare how horses run when they change claiming prices and conditions. I was going to do some research into, but this damned dizziness put a stop to that.:bang: I still think there might be a market for class numbers similar to Beyers for speed, but it just not going to be me who develops it.

The most useful approach I have found is in comparing how horses perform when "moved up" or "moved down" within adjacent class levels. That takes a mountain of coding to even begin to look at--but what it indicates (if successful) is that "class levels" are way different than ordinarily perceived.

I agree that there is probably a good market for class numbers (especially if they are isolated from, and do not depend on, pace and speed figures or conventional "par times").

[traynor]

I find that purse or earnings formulas are still useful in class determination, particularly if they also take consistency into account, and are supplemented in the proper proportion by other metrics addressing speed, running style, and condition.

I thought it was pretty well accepted "wisdom" that the inflated purses of state-bred races and racinos made/make use of the earnings box way less useful than 20-25 years ago. I generate all the standard values (in researching), but none seem really useful. In a limited number of races, it may seem that APV, EPS, win%, etc. are meaningful and significant. In larger collections, that significance seems to diminish.

[traynor]

Race purse, current and past.

Same problem as using the earnings box. Especially when the purses may be essentially the same for several different class levels (within or adjacent to other classes).

[thaskalos]

To me class is just one piece of the handicapping puzzle. In my own play...I am not trying to determine which horse qualifies to be called the "class of the race"; I am trying to determine if the horse "belongs" in the field which it is asked to face today.

Long and painful experience has convinced me that "class" cannot be accurately measured by speed and pace figures...nor can it be determined by a horse's earnings.

I am primarily a pace/speed handicapper...and I have found that most horses do not reproduce their lofty speed/pace figures when they are raised in class...even if this rise in class appears insignificant to the naked eye.

Consequently...I refuse to rate horses off of races which are lower "class-wise" than the races that they find themselves in today.

[traynor]

Classy horses are talented, and they have the qualities that are required to run the race on their own terms.

Talent is speed. Running the race on their own terms allows them to fully express their talent.

Class will negatively affect a "cheap" horse by way of adversity.''

I understand, but what exactly are you saying? I have read (and heard) endless explanations of why "cheap speed folds under pressure." The vast majority of the situations in which that appears to be the case are an illusion. The entry was (often) a one-race wonder that peaked and tailed off after a "good" race. The situation looks entirely different if studied closely, considering the probabilities of improved or declining performance regardless of the class level at which entered.

I have also read (and heard) endless explanations of how "class" manifests itself within a race to enable "higher class" horses to defeat "lower class" horses--while running slower than the "lower class" horses. It is nice in theory, but the reality seems to be based on a very limited sample of remembered events, rather than something that can be defined and quantified. It is like Janus' definition of groupthink--it can only be applied in retrospect.

[traynor]

Average earnings per start would be a good place to start.

Gerge Kaywood swore by EPS as a qualifier, especially at "lower class" tracks. Again, I am not so sure that it is as useful now as it was (or may have been) in the past.

[traynor]

The first thing I look at with any race is conditions and class ability of all the entrants with wins and up close finishes and as a thing in itself without any reference to speed or pace figures(at least initially).The location,surface,level,the movement around,it IS how you get to know the personality of the horse and the trainer.

What to do with it.I love outrun lines.It makes a horse shipping in from Hoosier Park to Beulah win at 4-1 or 10-1 because outrun speed competitive in today's race,instead of being in the vacuum of a faster race the horse finds itself in the ideal spot and gets aggressive.I love people who boast of never looking at past performances.It makes a horse that nearly won on OC62/N2X that is competing in an OC32/N1X win at 6-1.It makes layoff horse that nearly won a G3 race as a 2 year old 6 months ago win a N1X at 4-1 because it's speed figure is 6-8 points below the top in the field.Class is always relative to the field,but when comparing competitive speed the horse that earns it vs better is nearly always better than those who earn it vs lesser.I think only a few of us believe in it because it is such a race specific concept.It is only vaguely testable in a database.The latter being the criteria for belief,therefore it is not believed,therefore it is an overlay.

I'm not sure if I completely grasp where you were going with your post but this is the best answer I can give right now.Things like this arouse my interest because it fits neatly into my theory of all races as single factor events.1 factor that dominates a specific race and sits atop the hierarchy and renders secondary all other factors.And it can be so dominant that it can render 4-5 positive factors meaningless and underplayed.A single factor can shatter any theory of what is positive in any handicapping theory,in any database,single race dependant.Class is definitely one of those factors.And the more it is not believed the bigger it will indeed be overlayed.I like to look at it in isolation like I think you are saying you do.

Definitely. I agree wholeheartedly that speed vs better is more significant than speed vs lesser (or equal). That is what makes the study of race conditions and "real" class levels so useful--it generates values and insights that go right over the head of most bettors (including more than a few professionals).

For those who may be unfamiliar with the concept, a horse that is reasonably competive against a "better" field in a recent race (yet was soundly defeated) may actually be superior to a horse that scored a Big Win (or equivalent) against a lesser field. The winner looks great on paper but (as mentioned above) may simply be a one-race wonder that peaked.

[BlueChip@DRF]

Gerge Kaywood swore by EPS as a qualifier, especially at "lower class" tracks. Again, I am not so sure that it is as useful now as it was (or may have been) in the past.


I said it was just a good place to start, not the the final say. In general, the better races have higher purses.

[Overlay]

I thought it was pretty well accepted "wisdom" that the inflated purses of state-bred races and racinos made/make use of the earnings box way less useful than 20-25 years ago. I generate all the standard values (in researching), but none seem really useful. In a limited number of races, it may seem that APV, EPS, win%, etc. are meaningful and significant. In larger collections, that significance seems to diminish.
That "wisdom" about inflated purses was indeed (and apparently still is) "common". In fact, I recall Ainslie in the late '80's categorically saying that anyone who continued to rely on earnings data for class distinctions was "absolutely doomed" in jurisdictions where state-bred races with inflated purses were run. However, hard performance statistics (at least the ones I have seen) since then, plus my own experience over the past 25 years, tell me a different story. I stand by my previous post.

[MJC922]

I agree that there is probably a good market for class numbers (especially if they are isolated from, and do not depend on, pace and speed figures or conventional "par times").

Free data available through ADWs seems to have killed off a big chunk of the market. While I don't think there's much left in the way of people willing to pay anything 'extra' for this type of info I'd like to be wrong about that. For the caliber of folks on this board I do agree there's probably a market, but then I tend to think the folks on this board are far more serious students than most. I suspect the average person walking into a track has never visted this board even once. What a shame that is as this board opens the window to a rich experience of handicapping factors and different approaches making it clear even to a novice that he or she is now playing a superior gambling game.

[traynor]

To me class is just one piece of the handicapping puzzle. In my own play...I am not trying to determine which horse qualifies to be called the "class of the race"; I am trying to determine if the horse "belongs" in the field which it is asked to face today.

Long and painful experience has convinced me that "class" cannot be accurately measured by speed and pace figures...nor can it be determined by a horse's earnings.

I am primarily a pace/speed handicapper...and I have found that most horses do not reproduce their lofty speed/pace figures when they are raised in class...even if this rise in class appears insignificant to the naked eye.

Consequently...I refuse to rate horses off of races which are lower "class-wise" than the races that they find themselves in today.

I agree wholeheartedly that attempts to define "class" using earnings or pace and speed figures are not very useful. I have been doing this so long that I can remember when both John Meyers and Sartin advocated "adjusting" pace figures by 1/5 second for class moves up or down. That is, a horse was assumed to be somehow automatically capable of running a fifth of a second faster (at six furlongs)--if dropped one class level--based on a "pace line" from a "higher class" race. And a horse "moving up" in class was "penalized" a fifth of a second for each class level up.

It was a grand theory that--like many grand (and cherished) theories--was a complete flop in the real world.

I think many of the "moving up" horses that are defeated at the higher class are defeated because they are off-form, rather than that are uncompetitive at the higher class level. Some of the best bets one can encounter in "lower class" races are derived from tossing out the looks-good-on-paper-and-on-computer favorites that everyone seems to want to bet on.

[traynor]

That "wisdom" about inflated purses was indeed (and apparently still is) "common". In fact, I recall Ainslie in the late '80's categorically saying that anyone who continued to rely on earnings data for class distinctions was "absolutely doomed" in jurisdictions where state-bred races with inflated purses were run. However, hard performance statistics (at least the ones I have seen) since then, plus my own experience over the past 25 years, tell me a different story. I stand by my previous post.

I am asking a question, rather than trying to dispute the validity of anyone else's view. The stats I have seen (and generated) seem to indicate that evaluations based on earnings or conventional class levels are less than stellar in performance as predictive indicators of performance. YMMV.

[traynor]

Free data available through ADWs seems to have killed off a big chunk of the market. While I don't think there's much left in the way of people willing to pay anything 'extra' for this type of info I'd like to be wrong about that. For the caliber of folks on this board I do agree there's probably a market, but then I tend to think the folks on this board are far more serious students than most. I suspect the average person walking into a track has never visted this board even once. What a shame that is as this board opens the window to a rich experience of handicapping factors and different approaches making it clear even to a novice that he or she is now playing a superior gambling game.

I don't know about the average bettor, or about trying to market class ratings for a couple of dollars to anyone willing to buy them. I do think there is a definite market for such ratings for professional bettors, if the ratings are superior to (and more predictive than) currenlty available alternatives.

[traynor]

I said it was just a good place to start, not the the final say. In general, the better races have higher purses.

I should have qualified that. I have been (fairly steadily) monitoring all the "conventional" approaches to quantifying "class" for a number of years, including a number of proprietary (from other sources) algorithms derived from earnings in isolation and in combination with other factors. None--including EPS, APV, and a variety of others--have been particularly useful. I monitor them, research them, on and on--but for all practical purposes I would probably be better off to compelely ignore them.

The only class factors that I have found useful are those based on the relative probabilities of horses in good form being able to compete at other class levels, and then to use that information to build an alternative class level structure for a given track. Lots of work. I would really like to find an easier, simpler, less time-consuming way to quantify class that is as accurate (or more accurate) than what I am using now. That was my motivation to start this thread. I am open to any and all ideas, opinions, insights, conjectures, or whatever, and I have no desire to dispute anyone else's experience or findings.

However, earnings per start is one of the "conventional" approaches that I have found pretty much worthless for anything more than a number to look at on a computer screen or printout. Perhaps I am missing something, and there is some way to apply it or factor it in to the handicapping process that I am overlooking. I actually hope so, since it is one of a number of such figures that my app(s) generate for every entry in every race.

[Robert Fischer]

I understand, but what exactly are you saying?
That's asking for more than I'm willing to spell out, for the sake of stimulating discussion.

[thaskalos]

''

I understand, but what exactly are you saying? I have read (and heard) endless explanations of why "cheap speed folds under pressure." The vast majority of the situations in which that appears to be the case are an illusion. The entry was (often) a one-race wonder that peaked and tailed off after a "good" race. The situation looks entirely different if studied closely, considering the probabilities of improved or declining performance regardless of the class level at which entered.

I have also read (and heard) endless explanations of how "class" manifests itself within a race to enable "higher class" horses to defeat "lower class" horses--while running slower than the "lower class" horses. It is nice in theory, but the reality seems to be based on a very limited sample of remembered events, rather than something that can be defined and quantified. It is like Janus' definition of groupthink--it can only be applied in retrospect.
I have heard many explanations too...where people try to make sense of the inconsistency in the performance of claiming horses...as they go up and down the claiming ladder.

A $7,500 wins wire-to-wire while setting a fast pace. Three weeks later, this horse is entered in a $10,000 race, and it figures to be best again. No horse in this seemingly classier race looks capable of preventing our "cheaper" horse from repeating his impressive last-race victory. But our horse loses badly for $10,000...without even taking a clear lead. And it loses to a winner who ran slower than the cheaper horse ran in its previous, winning race.

What happened?

Did our horse lose because its form tailed off in its last race?

Doubtfull...because the horse often improves dramatically when it drops to its normal level next time.

Was the horse "intimidated" in its last race...by horses who threw super-quick mini-fractions at it...which could not be noticed by a casual glance at the pps?

I don't think that's what happened either. After all...how much difference could there be ability-wise between a razor-sharp $7,500 claimer...and its duller-looking $10,000 counterparts?

No...I think the difference between the two performances can be attributed to "trainer intent". The trainer is of the opinion that his $7,500 is "outclassed" in the higher race...and enters it strictly as preparation for the upcoming $7,500 claiming race...a couple of weeks down the line. You've seen a baseball batter swing a weighted bat as he is waiting his turn at the plate, right? Well...it's the same philosophy at work here.

That's why I feel sorry for those speed/pace handicappers who are of the opinion that class is a "non-factor" in handicapping.

Class is alive and well...even when it exists only in the trainer's mind.

[speculus]

My definition of CLASS in a horse is that ability in him which enables him to establish his superiority over rivals despite having a less than perfect trip.

Having said that, mathematically it is possible to get a grip on the elusive concept of CLASS if you have done some research about how a horse strides.

Though I do not wish to disclose too much, I don't mind giving a hint.

The whole world is obsessed with speed and pace, and hence they focus only on the distance (D) and time (t) to work out speed (S) which is nothing but distance divided by time.

S = D/t

But there is another way of looking at speed when you break it down to its basic components, which are to speed as atoms and molecules are to a physical object. These components are the basic constituents of speed.

These components of a horse's stride are his REACH and the TIME he takes to complete one stride.

REACH is the length of ground that a horse covers in one stride (the linear distance between two successive hoof marks made by the same leg of a galloping horse).

Let's denote REACH as R and the time taken for the stride as T.

Therefore, if a horse takes N strides to cover distance D in time t, then

R=D/N and T= t/N.

Thus you can see that speed can also be expressed as

S = R/T

Now, a hint: If one mathematical combination of R & T (ratio of R to T) gives us SPEED, one of the most important components of the racing puzzle, then can't some other combination of the same pair give us a clue to the horse's class?

How about R multiplied by T (R*T), or R squared divided by T (R^2/T) or R divided by T squared (R/T^2)? Or some other combination your creativity can think of? At this stage, it becomes a matter of research.

All I can say is that the key to mathematical assessment of the CLASS factor lies in this direction.

Go figure.;)

[traynor]

That's asking for more than I'm willing to spell out, for the sake of stimulating discussion.

I am not asking for anyone to give away the candy store--only for general information. I have seen (and helped develop) a number of algorithms that attempt to quantify "competitiveness" as a primary, with "class" as a secondary. Steve Chaplin, in the old Bettor's Guide for standardbreds, made an attempt at such quantification, and a number of people have continued along that line (and related lines). For many years.

One of the major issues in quantifying class is (as in many other aspects of racing) the "confounding variables." Meaning, the phenomena or factor observed may be contingent on, or the direct result of, some other (unknown or unobserved) variable factor. The classic example is a driving finish to the wire. One horse seems to thrive under the pressure and comes back stronger. The other "bounces." The "why" in many cases is the direct result of one or more confounding variables.

[Tom]

That's why I feel sorry for those speed/pace handicappers who are of the opinion that class is a "non-factor" in handicapping.

Class is alive and well...even when it exists only in the trainer's mind.

I think class is something you cannot put a number on for a horse. The pace and speed figures give you clues, and they can tell you what the operational class of the horse might be now, but class itself is more than that. You want to see class in action, watch Zenyatta in the last strides, refusing to lose. Put a number on that. Class was Affirmed not giving in to Alydara in the Belmont. It was Forego beating Honest Pleasure at the wire.

But I agree that what the trainer thinks class is is very playable every day.

[traynor]

I have heard many explanations too...where people try to make sense of the inconsistency in the performance of claiming horses...as they go up and down the claiming ladder.

A $7,500 wins wire-to-wire while setting a fast pace. Three weeks later, this horse is entered in a $10,000 race, and it figures to be best again. No horse in this seemingly classier race looks capable of preventing our "cheaper" horse from repeating his impressive last-race victory. But our horse loses badly for $10,000...without even taking a clear lead. And it loses to a winner who ran slower than the cheaper horse ran in its previous, winning race.

What happened?

Did our horse lose because its form tailed off in its last race?

Doubtfull...because the horse often improves dramatically when it drops to its normal level next time.

Was the horse "intimidated" in its last race...by horses who threw super-quick mini-fractions at it...which could not be noticed by a casual glance at the pps?

I don't think that's what happened either. After all...how much difference could there be ability-wise between a razor-sharp $7,500 claimer...and its duller-looking $10,000 counterparts?

No...I think the difference between the two performances can be attributed to "trainer intent". The trainer is of the opinion that his $7,500 is "outclassed" in the higher race...and enters it strictly as preparation for the upcoming $7,500 claiming race...a couple of weeks down the line. You've seen a baseball batter swing a weighted bat as he is waiting his turn at the plate, right? Well...it's the same philosophy at work here.

That's why I feel sorry for those speed/pace handicappers who are of the opinion that class is a "non-factor" in handicapping.

Class is alive and well...even when it exists only in the trainer's mind.

I have watched a lot of trainers use similar maneuvers, with varying degrees of success. Trainers often throw in a cooler race between serious races, using it as an extended workout. For claimers, it is normal to enter the horse at a higher claiming price to reap a bonus if claimed, or to avoid claims and obscure the form of a horse in good condition.

I agree that trainer intent is a major factor to consider. I have mentioned previously that (to me) one of the most useful techniques for understanding the realities of racing was one advocated by Bob Purdy, of the Sartin PIRCO group. That is to start with the oldest race in the PPs, and work up to the last race, with the sole thought in mind of trying to understand why the trainer entered the horse in each and every one of the races shown. Time consuming initially, but a good exercise to develop understanding.

In the scenario you describe, an even more interesting situation might develop if the entry wins when dropped back to $7500--and then is entered again at $10000--when everyone believes it has no chance to win.

[raybo]

I have watched a lot of trainers use similar maneuvers, with varying degrees of success. Trainers often throw in a cooler race between serious races, using it as an extended workout. For claimers, it is normal to enter the horse at a higher claiming price to reap a bonus if claimed, or to avoid claims and obscure the form of a horse in good condition.

I agree that trainer intent is a major factor to consider. I have mentioned previously that (to me) one of the most useful techniques for understanding the realities of racing was one advocated by Bob Purdy, of the Sartin PIRCO group. That is to start with the oldest race in the PPs, and work up to the last race, with the sole thought in mind of trying to understand why the trainer entered the horse in each and every one of the races shown. Time consuming initially, but a good exercise to develop understanding.

In the scenario you describe, an even more interesting situation might develop if the entry wins when dropped back to $7500--and then is entered again at $10000--when everyone believes it has no chance to win.

The method of starting with the oldest race and working forward is something I have done, and advocated on this forum, for years. By doing this you gain insight into class, distance capability, form cycles and current form, and trainer intent. And, there is great value in doing this, as very few players are willing to take the time, or to put forth the effort to accomplish it. It's one of many "value" approaches available to the player who isn't afraid of a little work.

[thaskalos]

I have watched a lot of trainers use similar maneuvers, with varying degrees of success. Trainers often throw in a cooler race between serious races, using it as an extended workout. For claimers, it is normal to enter the horse at a higher claiming price to reap a bonus if claimed, or to avoid claims and obscure the form of a horse in good condition.

I agree that trainer intent is a major factor to consider. I have mentioned previously that (to me) one of the most useful techniques for understanding the realities of racing was one advocated by Bob Purdy, of the Sartin PIRCO group. That is to start with the oldest race in the PPs, and work up to the last race, with the sole thought in mind of trying to understand why the trainer entered the horse in each and every one of the races shown. Time consuming initially, but a good exercise to develop understanding.

In the scenario you describe, an even more interesting situation might develop if the entry wins when dropped back to $7500--and then is entered again at $10000--when everyone believes it has no chance to win.

I have been doing this for at least 15 years...ever since Henry Kuck recommended it in his book WINNER'S FILE.

I don't remember anything else from that book...but that's a piece of advice that I have never forgotten.

[MJC922]

I think class is something you cannot put a number on for a horse. The pace and speed figures give you clues, and they can tell you what the operational class of the horse might be now, but class itself is more than that. You want to see class in action, watch Zenyatta in the last strides, refusing to lose. Put a number on that. Class was Affirmed not giving in to Alydara in the Belmont. It was Forego beating Honest Pleasure at the wire.

But I agree that what the trainer thinks class is is very playable every day.

If a horse is actually quickening we could put a number on it provided that we had data coming from a realtime source. The trouble is what we have is (at best) a 6+ second internal split running into a headwind or tailwind which itself would also require real time data to put it into context. Yeah I believe some horses are smart and win close ones, Forty Niner, Favorite Trick etc, but there are also others without a whole lot of intangibles who would wipe the floor with the former e.g. Skip Away.

[Robert Fischer]

Two different horses can be entered in the same race, and within that same race, their class can be tested by far different levels of adversity. This can be estimated before the race and calculated after the race.

[Capper Al]

I don't know about the average bettor, or about trying to market class ratings for a couple of dollars to anyone willing to buy them. I do think there is a definite market for such ratings for professional bettors, if the ratings are superior to (and more predictive than) currenlty available alternatives.

I have thought about selling class figures like CJ sells speed figurers here. I'm happy with what I have and they actually outperform my speed figs, but my speed figs are under revision. Mine are based on BRIS' RR and CR ratings. The RR is used for rating class in the traditional sense as most cappers would think of it. The CR is used for that mixed relationship with form. As to what to do with the class/speed relationship, I believe those authors (like Ainslie and Quirin) give good examples of how to do this.

There are two main beliefs about class and, therefore, two fundamental methods of measuring class. The oldest approach, my choice, is based on race performance within race type. And yes, earnings is at the heart of this measure. The other, the modern, is speed which is measured via par times for race type. I prefer the race performance method because it gets around speed figs which won't explain how the speed was made.

[traynor]

I have thought about selling class figures like CJ sells speed figurers here. I'm happy with what I have and they actually outperform my speed figs, but my speed figs are under revision. Mine are based on BRIS' RR and CR ratings. The RR is used for rating class in the traditional sense as most cappers would think of it. The CR is used for that mixed relationship with form. As to what to do with the class/speed relationship, I believe those authors (like Ainslie and Quirin) give good examples of how to do this.

There are two main beliefs about class and, therefore, two fundamental methods of measuring class. The oldest approach, my choice, is based on race performance within race type. And yes, earnings is at the heart of this measure. The other, the modern, is speed which is measured via par times for race type. I prefer the race performance method because it gets around speed figs which won't explain how the speed was made.

From your description it sounds like you are pretty much following the same pathway that others have taken in attempting to quantify class. I think that the most useful (and most profitable) would be to come up with a different paradigm for quantifying class that does not rely on earnings or "class/speed pars." Both those approaches have been hammered into oblivion by anyone with access to a computer.

Basically, because it is easy to affix what seem to be "meaningful numbers" to a horse and declare it a measure of class, a LOT of computer handicappers dabble around with various combinations and approaches. I don't think it is that simple (or that easy). That is not a spurious statement--I have done a number of years of very serious study of the topic, and--as a consequence--tossed most conventional approaches to quantifying class as "interesting, but not especially useful."

The reason is much like speed and pace figures--readily available to anyone with a computer. It is not rocket science to program a few simple algorithms to massage earnings, performance, and pars and call the result a measurer of "class." Again, I don't think it is that simple (or that easy).

The old saying that, "To a carpenter with a hammer, every problem looks like a nail" is equally applicable to computer users. There is a tacit belief that the best solution to every problem in horse racing is in massaging the same set of numbers downloaded every day by every other computer handicapper out there. That is rarely the case.

I think Robert Fischer is on the right track. He alludes to insights that are being used actively (and currently) by some very serious bettors. Whether those insights are of the same phenomena or not, I have no way of knowing. It may well be that the best quantification of class will be a combination of competitiveness and in-depth study of race conditions and real world class levels with earnings, performance, and pars.

[traynor]

Two different horses can be entered in the same race, and within that same race, their class can be tested by far different levels of adversity. This can be estimated before the race and calculated after the race.

Yes. In a dozen or more different ways.

[PhantomOnTour]

Class is like pornography - hard to specifically define, but you know it when you see it.

[traynor]

I have been doing this for at least 15 years...ever since Henry Kuck recommended it in his book WINNER'S FILE.

I don't remember anything else from that book...but that's a piece of advice that I have never forgotten.

I think too many confuse "staring at past performances" with "studying past performances." Such exercises go a long way in focusing attention on the critical factors displayed in past performances. I was fortunate to have attended numerous Sartin seminars in the 1980s at which Bob Purdy was a major contributor. The process was even more useful for those who took the time to run pace of race/pace of horse figures for each race studied. Entering the data for each race into a handheld Sharp 1350 was tedious, but created a whole new way of looking at past performances.

(For those who may not be familiar with the POR/POH calculations, it essentially generates a value (percentile) indicating how close a horse was to the leader at each positional call in the race. Increasing values indicate--explicitly--a horse "running closer to the pace" in subsequent race. Just one of the many patterns studied by the Sartin users who were actually betting on that stuff. Mike Fiore, in Miami, was advocating a similar approach about the same time, using lengths behind and "speed carried farther.")

[traynor]

Class is like pornography - hard to specifically define, but you know it when you see it.

I agree. That is the problem. Again, like Janus' definition of groupthink, it can only be categorized as being such after the fact. That tends to diminish the value when it is necessary to bet on it before you see it.

It is easy to declare that class enabled thus and so to happen in a given race. It is more difficult to declare before the race that thus and so is going to happen in that race because of class. Add to that the human tendency to selectively remember events that validate preconceived notions while conveniently forgetting events that provide overwhelming evidence to the contrary, and it becomes clear how useful it is to study a concept as nebulous as class.

[Capper Al]

From your description it sounds like you are pretty much following the same pathway that others have taken in attempting to quantify class. I think that the most useful (and most profitable) would be to come up with a different paradigm for quantifying class that does not rely on earnings or "class/speed pars." Both those approaches have been hammered into oblivion by anyone with access to a computer.

Basically, because it is easy to affix what seem to be "meaningful numbers" to a horse and declare it a measure of class, a LOT of computer handicappers dabble around with various combinations and approaches. I don't think it is that simple (or that easy). That is not a spurious statement--I have done a number of years of very serious study of the topic, and--as a consequence--tossed most conventional approaches to quantifying class as "interesting, but not especially useful."

The reason is much like speed and pace figures--readily available to anyone with a computer. It is not rocket science to program a few simple algorithms to massage earnings, performance, and pars and call the result a measurer of "class." Again, I don't think it is that simple (or that easy).

The old saying that, "To a carpenter with a hammer, every problem looks like a nail" is equally applicable to computer users. There is a tacit belief that the best solution to every problem in horse racing is in massaging the same set of numbers downloaded every day by every other computer handicapper out there. That is rarely the case.

I think Robert Fischer is on the right track. He alludes to insights that are being used actively (and currently) by some very serious bettors. Whether those insights are of the same phenomena or not, I have no way of knowing. It may well be that the best quantification of class will be a combination of competitiveness and in-depth study of race conditions and real world class levels with earnings, performance, and pars.

Sounds like to me that you think you already know the answer. Might be some superiority complex on your part. What didn't you understand about me being happy with my class figs?

[DeltaLover]

Class is with no doubt one of the most mysterious handicapping attributes.

As a novice is learning the game, very soon he discovers the concept of class.

In contrary to things like speed figures, track variants or impact values nobody seems to be in position to provide a clear and comprehensive explanation of how to measure it!

Based in this it becomes easy for the novice, to go ahead and ignore class as a factor all together. This approach will soon create quite a few incomprehensible results crying for a systematic error. Bayer talks about such a case in one of his books, where his speed figure based cinch, was crashed by the odds on favorite of the race who although much inferior speed wise, verified his class superiority in the race.

We all have similar experiences, the most common of them, is about a horse dropping from MSW to MCL and although been no factor any time before, now he romps to an easy win, some times at hefty odds!

So, there must be something called 'class', but what exactly it is?

How can we quantify it?

Or we just cannot?

The three main components of what we call class are the following:

(1) What the connections of the horse think about its ability as a runner and what are their intentions for entering him in a specific event

(2) What the crowd thinks about the ability of the runner

(3) The measurable ability of the runner

Of course there is a very high degree of correlation among (3) and the other two and as bettors what is the most important is the discrepancies between (2) and (3).

To understand 'class' we need to realize that the less accurate we can be about (3) the more profitable situations can be presented.

A 5 years old horse who's career consists of 20 starts and 3 wins has very little probability to confuse either his connections of the crowd as far his 'class'.

In contrary a first time out of a very lightly raced horse is quite the opposite and is where the larger errors in class estimate occur (both from the connections and the crowd).

When a runner starts his career we have very few data to classify him. Both (1) and (2) have a great margin to get confused. What is important to realize here is that any race starts as the reflection of the opinions and intentions of the connections. Some times they will be proven correct in their estimates but quite often the opposite will be true.

Of course they will be correct more often than not and this is why race classification is a valid metric for the quality of the race.

Based on this a MSW usually will be tougher than a MCL so when a horse who failed in it goes to a MCL even most likely will have an easier task facing inferior opponents. The crowd seems to be well aware about these obvious class moves and usually incorporates them accordingly to the betting lines.

Besides that, there are always exceptions to the rule, cases where the crowd is confused either due to a mistake of the trainer or because the trainer managed to deceive it creating wrong impressions about the capacity of his runner. These are the best opportunities for the astute handicapper to find a good bet.

As far as 'class figures', I do not think they will add any value to a handicapper's arsenal, since as speed figures, measurable class is reflected to the pools to a very high degree. I think that the final judgment of the 'class' of horse in comparison to his rivals represents the most obscure characteristic of handicapping and can only be attacked by intuition and judgment calls.

Class is one of the two deepest topics (form been the other) and we can talk about it for very long time without reaching a unanimous verdict since they both are more tangible to the artistic rather than the scientific part of handicapping...

[Ocala Mike]

Class is like pornography - hard to specifically define, but you know it when you see it.



This is so true. Actually, the horses probably have a better take on how to define class than we do. All you have to do is watch a group of mares out in the field or a group of babies playing around and you will see that they quickly sort themselves out.

[raybo]

As has been stated over and over and over, class is multifaceted. And, means something different to different players. There is what I call "true" class, and there is "calculated" class, or "performance" class.

"True" class is inherent to the animal, it's psychological makeup towards competitiveness, or in the wild, the "leader of the pack" level in which the horse belongs (herd stuff).

"Calculated" or "Performance" class is an attempt to quantify class by man-made class heirarchy and/or race conditions, and/or pace versus speed calculations, etc..

"True" class must be visually observed, for the most part, an observation of a horse that refuses to give up, for example, regardless of his/her available/remaining energy levels. Or, the ability to intimidate weaker or more "cowardly" horses (again, all herd stuff)

"Calculated" class, is just that, calculated from past race data and does not necessarily require visual observation.

By combining the 2 types of class one attains a "composite" class, which is probably the most valuable way of determining a class rating. But, it is also the most difficult to attain as most players do not visually observe and record "true" class, and also don't put enough thought into their "calculated" class ratings.

Personally, I prefer to use "calculated" class, and combine that with my velocities/energy/total velocity ratings and running styles. I feel that the "visual" part of the equation is far too subjective in nature, and therefore will not stand the test of time.

[Magister Ludi]

As has been stated over and over and over, class is multifaceted. And, means something different to different players. There is what I call "true" class, and there is "calculated" class, or "performance" class.

"True" class is inherent to the animal, it's psychological makeup towards competitiveness, or in the wild, the "leader of the pack" level in which the horse belongs (herd stuff).

"Calculated" or "Performance" class is an attempt to quantify class by man-made class heirarchy and/or race conditions, and/or pace versus speed calculations, etc..

"True" class must be visually observed, for the most part, an observation of a horse that refuses to give up, for example, regardless of his/her available/remaining energy levels. Or, the ability to intimidate weaker or more "cowardly" horses (again, all herd stuff)

"Calculated" class, is just that, calculated from past race data and does not necessarily require visual observation.

By combining the 2 types of class one attains a "composite" class, which is probably the most valuable way of determining a class rating. But, it is also the most difficult to attain as most players do not visually observe and record "true" class, and also don't put enough thought into their "calculated" class ratings.

Personally, I prefer to use "calculated" class, and combine that with my velocities/energy/total velocity ratings and running styles. I feel that the "visual" part of the equation is far too subjective in nature, and therefore will not stand the test of time.

Would you agree that a succinct definition of class would be “the intangible and nonrandom component that remains after a horse’s performance is decomposed into speed, pace, form, jockey, trainer, and other quantifiable factors?”

[raybo]

Would you agree that a succinct definition of class would be “the intangible and nonrandom component that remains after a horse’s performance is decomposed into speed, pace, form, jockey, trainer, and other quantifiable factors?”

Yes I agree, that would be "true" class, or "inherent" class. However, I think it too is "quantifiable", or put into a number/letter etc., format.

[pondman]

Would you agree that a succinct definition of class would be “the intangible and nonrandom component that remains after a horse’s performance is decomposed into speed, pace, form, jockey, trainer, and other quantifiable factors?”

No.

Why insist that these variables are part of a class equation or function? And need to be peeled away prior to finding something hidden?

Class stands by itself.
You can routinely bet on it without PPs.

[raybo]

No.

Why insist that these variables are part of a class equation or function? And need to be peeled away prior to finding something hidden?

Class stands by itself.
You can routinely bet on it without PPs.

I think you are misunderstanding what he is saying. He is saying that "inherent"/"true" class does stand by itself. The other factors he mentioned are just that, other factors. If you determine class by including other factors then it becomes a "composite" or calculated class, ie: a "performance" class rating.

[FiveWide]

Does anyone factor a horse's age into their class equation?


-Five

[raybo]

Does anyone factor a horse's age into their class equation?


-Five

IMO, age should be factored into all ratings.

[DeltaLover]

IMO, age should be factored into all ratings.
What matters the most is not that much the age of the horse as how many races he has ran so far. The less races the higher the probability for a wrong class estimate (either by crowd or connections) which is what we are looking for as handicappers and bettors.

[pondman]

I think you are misunderstanding what he is saying. He is saying that "inherent"/"true" class does stand by itself. The other factors he mentioned are just that, other factors. If you determine class by including other factors then it becomes a "composite" or calculated class, ie: a "performance" class rating.

I don't like the word "remain" in the statement. I also don't like the word "intangible", as if class is a form of good or bad will.

[pondman]

What matters the most is not that much the age of the horse as how many races he has ran so far. The less races the higher the probability for a wrong class estimate (either by crowd or connections) which is what we are looking for as handicappers and bettors.

Or the correct assessment, if the horse is a 2 or 3 year old. I believe this to be a much more profitable end for those who play the young horses, than playing the older horse scenario.

[raybo]

What matters the most is not that much the age of the horse as how many races he has ran so far. The less races the higher the probability for a wrong class estimate (either by crowd or connections) which is what we are looking for as handicappers and bettors.

I agree, but there is still an "age" component, as young horses mature, in every aspect, as they age, and in some aspects regardless of the number of races run.

[raybo]

I don't like the word "remain" in the statement. I also don't like the word "intangible", as if class is a form of good or bad will.

I think "remain" directs towards their base class, what they were born with. "Intangible" has nothing to do with "good or bad", it simply means something not tangible, as in it can't necessarily be measured, in a traditional mathematical/material way.

[traynor]

Sounds like to me that you think you already know the answer. Might be some superiority complex on your part. What didn't you understand about me being happy with my class figs?

Because you are happy with them does not indicate that anyone else would be willing to pay for them. Of course, that is easy to determine. Offer them for sale and see how many takers you get.

[raybo]

Because you are happy with them does not indicate that anyone else would be willing to pay for them. Of course, that is easy to determine. Offer them for sale and see how many takers you get.

Oh he'll get some takers, for sure. But, probably not much return business. :lol:

[traynor]

A major part of the problem may be that "class" means something different to each person. Another big part may be that "stuff" (mostly unforeseen results) is attributed generically to the effects of "class" when it may be the result of entirely different factors/circumstances.

In DeltaLover's example of MSW dropping to MCLM, there may be far more going on than a "higher class" horse dropping into a "lower class" race. That is the surface. Beneath that surface is the reality that the perception of a horse as being allowance/stakes grade has changed to being just another claimer--for sale to anyone with the price (qualified to buy, of course). It may be that the trainer/stable/owner is taking great care to assure at least a win (and winning wagers) if the horse is claimed. Extra effort may be made that has an influence on the race outcome--including more careful selection of the field against which the horse will compete, with late scratches against tough fields.

It may also be that the trainer/stable/owner has been using MSW races to condition a horse that is assumed to be "just another claimer" in the hopes of selling it for a premium price when "dropped in class." This is a business, folks. Blurting out one's strategies to anyone who will listen is not usually a part of what trainers do.

With that in mind, the MSW winner in a MCLM (and many, many similar situations) may have little or nothing to do with "class," and trying to "explain" something that is part of a trainer's business strategy may do little more than further obscure an already obscure phenomena.

[raybo]

A major part of the problem may be that "class" means something different to each person. Another big part may be that "stuff" (mostly unforeseen results) is attributed generically to the effects of "class" when it may be the result of entirely different factors/circumstances.

In DeltaLover's example of MSW dropping to MCLM, there may be far more going on than a "higher class" horse dropping into a "lower class" race. That is the surface. Beneath that surface is the reality that the perception of a horse as being allowance/stakes grade has changed to being just another claimer--for sale to anyone with the price (qualified to buy, of course). It may be that the trainer/stable/owner is taking great care to assure at least a win (and winning wagers) if the horse is claimed. Extra effort may be made that has an influence on the race outcome--including more careful selection of the field against which the horse will compete, with late scratches against tough fields.

It may also be that the trainer/stable/owner has been using MSW races to condition a horse that is assumed to be "just another claimer" in the hopes of selling it for a premium price when "dropped in class." This is a business, folks. Blurting out one's strategies to anyone who will listen is not usually a part of what trainers do.

With that in mind, the MSW winner in a MCLM (and many, many similar situations) may have little or nothing to do with "class," and trying to "explain" something that is part of a trainer's business strategy may do little more than further obscure an already obscure phenomena.

Of course!

[traynor]

Back in the 1980s, Michael Pizzola did a presentation at a Sartin seminar in New Orleans. Pizzola described a strategy used by a New York betting syndicate that combined class evaluations (using a proprietary process that was unrelated to the Sartin apps) with a specific money management strategy. The crux of the matter was that the class evaluations that formed the basis of the strategy selected enough winners at high enough prices to shift the focus from "just winning" to optimizing return. It was a fascinating presentation.

Even more fascinating was the number of people who shifted their focus in race analysis from the (generally accepted as The Way and The Light) emphasis on pace and speed to class. That emphasis was especially evident in the Sartin users, in particular when James Quinn "became associated" with the Sartin group, and in-depth understanding of The Handicapper's Condition Book, class levels, and the overall effect of class on race outcomes was considered entry-level expertise.

That was like 25-30 years ago. And now--all those many years later--people still believe that massaging earnings, purse levels, and pars are going to give them some kind of "edge" that no one else has. Been there. Done that. Along with many, many others. I think it takes a bit more to gain an edge in effectively using class than massaging a few readily available (to everyone) numbers and declaring that one has discovered the greatest thing since sliced bread.

That is not intended in any way to criticize anyone's definitions or strategies for analyzing class. It is only intended to state that a lot of highly motivated people have been working on those strategies for many years, and most have discarded the use of earnings, purse levels, and pars as insufficient to be of real value for serious bettors. Of course, that is only my own (highly subjective) opinion. YMMV.

[classhandicapper]

I determine class by measuring the quality of the "fields" a horse has been competing in and then comparing how that horse did within those fields compared to other horses given their trips (pace, bias, ground, trouble etc...).

You measure the quality of the fields by developing an understanding of the pecking order at your track (circuit) and examining exactly who was in the race and how well they ran. It also helps if you know the common shipper tracks.

[raybo]

Back in the 1980s, Michael Pizzola did a presentation at a Sartin seminar in New Orleans. Pizzola described a strategy used by a New York betting syndicate that combined class evaluations (using a proprietary process that was unrelated to the Sartin apps) with a specific money management strategy. The crux of the matter was that the class evaluations that formed the basis of the strategy selected enough winners at high enough prices to shift the focus from "just winning" to optimizing return. It was a fascinating presentation.

Even more fascinating was the number of people who shifted their focus in race analysis from the (generally accepted as The Way and The Light) emphasis on pace and speed to class. That emphasis was especially evident in the Sartin users, in particular when James Quinn "became associated" with the Sartin group, and in-depth understanding of The Handicapper's Condition Book, class levels, and the overall effect of class on race outcomes was considered entry-level expertise.

That was like 25-30 years ago. And now--all those many years later--people still believe that massaging earnings, purse levels, and pars are going to give them some kind of "edge" that no one else has. Been there. Done that. Along with many, many others. I think it takes a bit more to gain an edge in effectively using class than massaging a few readily available (to everyone) numbers and declaring that one has discovered the greatest thing since sliced bread.

That is not intended in any way to criticize anyone's definitions or strategies for analyzing class. It is only intended to state that a lot of highly motivated people have been working on those strategies for many years, and most have discarded the use of earnings, purse levels, and pars as insufficient to be of real value for serious bettors. Of course, that is only my own (highly subjective) opinion. YMMV.

Pizzola has always emphasized value over picking winners. The fact that he presented a "class" example, was probably here nor there. The same principal would apply to any other method; pace, speed, etc.. The reason why many switched from pace and speed to class, subsequently, was probably because they didn't "get" the primary point of the demonstration, that being choosing value, rather than just picking winners.

[raybo]

I determine class by measuring the quality of the "fields" a horse has been competing in and then comparing how that horse did within those fields given the trips (pace, bias, ground, trouble etc...).

You measure the quality of the fields by developing an understanding of the pecking order at your track (circuit) and examining exactly who was in the race and how well they ran.

And how do you determine the "pecking order"; class, wins, pace/speed, etc.? That still doesn't express how you define class.

[Capper Al]

Because you are happy with them does not indicate that anyone else would be willing to pay for them. Of course, that is easy to determine. Offer them for sale and see how many takers you get.

You won't let that know it all attitude go, will you. I didn't say they were for sale. My top two class figs will produce the winner over 50%, and they are as imdependent from speed as EPS are. That is something to be happy about.

[RaceBookJoe]

When I first started handicapping, I used class ratings, but the problem I always had was that class wasn't static. Class is pace and speed, but its also a level where a trainer might feel the horse is competitive, and its also current form, it can even be a change of track maneuver etc etc. So with all of that, I just did my best both trying to assess where the horse was form-wise and what was going on inside the trainers mind for the race I was handicapping. Very complicated subject to say the least.

[classhandicapper]

And how do you determine the "pecking order"; class, wins, pace/speed, etc.? That still doesn't express how you define class.

I believe the race conditions are "generally" efficient. That is, a 50K claiming field in NY will generally be superior to a 25K claimer overall etc.... However, there will be some overlap, weak fields, strong fields at each level.

For example, an extremely impressive winner at the 25K level may be better than the typical 50K horse and some 50K fields will be made up a bunch of horses that have been getting buried at that level and are actually much weaker etc... So you have to dig into the fields.

You can get a general gauge on the pecking order by looking at how horses do when they move up and down the scale and what moves produce winners and what moves produce failures.

You can look at race PARs to get a feel for how fast certain levels are "on average" and then look at the makeup of a specific field to determine if it is above or below average.

You can follow horses out of races to see how they do next out.

You can look for common opponents.

etc....

The more of this you do, the better you will understand the pecking order.

[traynor]

I determine class by measuring the quality of the "fields" a horse has been competing in and then comparing how that horse did within those fields compared to other horses given their trips (pace, bias, ground, trouble etc...).

You measure the quality of the fields by developing an understanding of the pecking order at your track (circuit) and examining exactly who was in the race and how well they ran. It also helps if you know the common shipper tracks.

That is very much like what I am doing currently (and have been working on for the last year or so). I am finding numerous cases in which a given "class" of race (based on specific race conditions) can be associated with the relative strength of the field in that race to create what is essentially a "class projection" for the race. Where it gets interesting is the situations in which a horse is entered (and competes in) a race at one class level that is "weaker" than races at a "lower" class level, or the reverse. Most bettors seem to wager according to relatively simplistic notions of class level and the competition a horse faces at those class levels. Lots of opportunities for anyone willing to do a bit of work and serious study.

[traynor]

I have thought about selling class figures like CJ sells speed figurers here. I'm happy with what I have and they actually outperform my speed figs, but my speed figs are under revision. Mine are based on BRIS' RR and CR ratings. The RR is used for rating class in the traditional sense as most cappers would think of it. The CR is used for that mixed relationship with form. As to what to do with the class/speed relationship, I believe those authors (like Ainslie and Quirin) give good examples of how to do this.

There are two main beliefs about class and, therefore, two fundamental methods of measuring class. The oldest approach, my choice, is based on race performance within race type. And yes, earnings is at the heart of this measure. The other, the modern, is speed which is measured via par times for race type. I prefer the race performance method because it gets around speed figs which won't explain how the speed was made.

My mistake. I thought it was a sales pitch.

[Ocala Mike]

Back in the day, I seem to recall reading a handicapping book where the author gave a very simple answer to the original question. Of course, this would only work for claiming races, which I guess were more plentiful then than they are now.

Take the HIGHEST claiming price that the horse has WON FOR in the PP's plus the LOWEST claiming price that the horse has RUN FOR in the PP's and divide by 2. Voila, you have a class value!

P.S., I never found this very successful or profitable.

[MJC922]

That is very much like what I am doing currently (and have been working on for the last year or so). I am finding numerous cases in which a given "class" of race (based on specific race conditions) can be associated with the relative strength of the field in that race to create what is essentially a "class projection" for the race. Where it gets interesting is the situations in which a horse is entered (and competes in) a race at one class level that is "weaker" than races at a "lower" class level, or the reverse. Most bettors seem to wager according to relatively simplistic notions of class level and the competition a horse faces at those class levels. Lots of opportunities for anyone willing to do a bit of work and serious study.

In my opinion you and classhandicapper are going about it the right way, it's about digging deeper into the field makeup. Labels, claiming prices, one dimensonal stuff like that all shows up on the board. FWIW, the class ratings that I generate are around for those that happen to stuble upon them. Since I'm using them myself if I were to market them to 'professional bettor's' well that would probably not be in my best interest anyway. Even though time is not part of the calculations, nor are earnings, or pars or any of that downstream stuff I have to suspect they still aren't far superior to any good solid speed figures. With that being said I also doubt they're inferior in measuring a horse's effort in any given race either, and far less distributed. To be clear to the board admin, I appreciate the exchange of ideas on the board and that's as far as it goes. In the end my ratings are just another tool, I need a way to classify and quantify efforts, but the key to making any money in the game is who is going to run a peak race today, unfortunately class has very little to do with who feels like running their eyeballs out, the latter is a better bet than the class of the field every time.

[classhandicapper]

That is very much like what I am doing currently (and have been working on for the last year or so). I am finding numerous cases in which a given "class" of race (based on specific race conditions) can be associated with the relative strength of the field in that race to create what is essentially a "class projection" for the race. Where it gets interesting is the situations in which a horse is entered (and competes in) a race at one class level that is "weaker" than races at a "lower" class level, or the reverse. Most bettors seem to wager according to relatively simplistic notions of class level and the competition a horse faces at those class levels. Lots of opportunities for anyone willing to do a bit of work and serious study.

This is a big part of my game. It takes a lot of work and is not easily packaged into a number. That's why there is value in it when done well.

[traynor]

This is a big part of my game. It takes a lot of work and is not easily packaged into a number. That's why there is value in it when done well.

If it was easy, every kid on the block would be doing it, and it would have no value. The further I go, the more I realize how much an advantage it is that a lot of handicappers are stuck back in the old days of speed numbers, simplistic pace analysis, and recycled strategies that have not changed much since the days of Ainslie, Beyer, Quirin, Sartin, et al. Or making great discoveries based on samples of a few hundred races.

It makes the work more rewarding for those willing to do it.

[CincyHorseplayer]

Back in the 1980s, Michael Pizzola did a presentation at a Sartin seminar in New Orleans. Pizzola described a strategy used by a New York betting syndicate that combined class evaluations (using a proprietary process that was unrelated to the Sartin apps) with a specific money management (#) strategy. The crux of the matter was that the class evaluations that formed the basis of the strategy selected enough winners at high enough prices to shift the focus from "just winning" to optimizing return. It was a fascinating presentation.

Even more fascinating was the number of people who shifted their focus in race analysis from the (generally accepted as The Way and The Light) emphasis on pace and speed to class (#). That emphasis was especially evident in the Sartin users, in particular when James Quinn "became associated" with the Sartin group, and in-depth understanding of The Handicapper's Condition Book, class levels, and the overall effect of class on race outcomes was considered entry-level expertise.

That was like 25-30 years ago. And now--all those many years later--people still believe that massaging earnings, purse levels, and pars are going to give them some kind of "edge" that no one else has. Been there. Done that. Along with many, many others. I think it takes a bit more to gain an edge in effectively using class than massaging a few readily available (to everyone) numbers and declaring that one has discovered the greatest thing since sliced bread.

That is not intended in any way to criticize anyone's definitions or strategies for analyzing class. It is only intended to state that a lot of highly motivated people have been working on those strategies for many years, and most have discarded the use of earnings, purse levels, and pars as insufficient to be of real value for serious bettors. Of course, that is only my own (highly subjective) opinion. YMMV.

This is a great post.The only way to get to know a horse(outside of total visual accessibility) is to look at running lines in relation to class levels,with every up and down tendency of form cycle you know throw in as a possibility,to actually get a genuine feel for the personality of the horse,what the connections infer about him or her based on their movement,and only then can a class assessment be made about the horse.Then all the objeective work be done in relating that to the field.As you and others have mentioned, over time scanning PP's bottom to top to get this feel is not burdensome and tedious.For me it instills the feeling of confidence in my assessment of horses.It determines the big or little of my bets.And the further we get into the computer age the bigger the overlay it becomes as past performances are proudly thrown out as yesterday's tools.It amazes me at times.I look at horses ,uch the same way as I look at boxers.There is no technological advancement that renders useless a jab,left hook,right hand,or in boxer's terms for class=ring generalship and the ability to take a punch.As much as I have become specific in my playing forte my approach is always meant to be universally applicable,and mainly because I like to play different tracks on a regular basis.Taking things like class assessment and looking at running lines and an array of pace figures(not compound rankings),these things make me feel as close as humanly possible to seeing into the raw reality of this game.Sorry for the long winded post,but this conversation stimulates me in so many ways.Awesome.

[Maximillion]

This is a great post.The only way to get to know a horse(outside of total visual accessibility) is to look at running lines in relation to class levels,with every up and down tendency of form cycle you know throw in as a possibility,to actually get a genuine feel for the personality of the horse,what the connections infer about him or her based on their movement,and only then can a class assessment be made about the horse.Then all the objeective work be done in relating that to the field.As you and others have mentioned, over time scanning PP's bottom to top to get this feel is not burdensome and tedious.For me it instills the feeling of confidence in my assessment of horses.It determines the big or little of my bets.And the further we get into the computer age the bigger the overlay it becomes as past performances are proudly thrown out as yesterday's tools.It amazes me at times.I look at horses ,uch the same way as I look at boxers.There is no technological advancement that renders useless a jab,left hook,right hand,or in boxer's terms for class=ring generalship and the ability to take a punch.As much as I have become specific in my playing forte my approach is always meant to be universally applicable,and mainly because I like to play different tracks on a regular basis.Taking things like class assessment and looking at running lines and an array of pace figures(not compound rankings),these things make me feel as close as humanly possible to seeing into the raw reality of this game.Sorry for the long winded post,but this conversation stimulates me in so many ways.Awesome.

This is what I used to do,but now I like to look at last 3, or last "good" race as a starting point.It saves some time without (hopefully) compromising getting a feel for what the horse can do.

[JohnGalt1]

I believe that a numerical rating can be given for class as well as pace/speed. Because I make one and it works for me, especially in route and turf races.

When I handicap I use a variation of the Performance Class Rating (PRC) from William L. Scott's "Total Victory at the track.)

To make the task simpler I created a comparison chart of class levels, though nothing is set in stone.

Class levels are in 20% increments higher and lower.

A $6,250 Stakes race is equivalent to a $8k Alw and a $10k starter Alw or OpClm and $16k open claiming and a $25k Md Spc and a $80k mdn Clm.

My chart runs from 1==Grade 1 races to 30 $1k claiming and $3200 Mdn clm.

I adjust one level down for NW3L, and two levels down for state bred and NW2L. Example a 50k state bred NW2L would be equivalent to an open $20k claiming race or a 32k mdn special race.

And I might adjust for track to track. A track with 25k claimers running for $30k as against a $25 claimer running at a different track for $13k might have an adjustment.

Then I make my pace/class figures and if a horse is one of the top horses in both categories and is in form, then I have a strong contender.

I don't have rules when I handicap. I have guidelines.

[traynor]

This is a great post.The only way to get to know a horse(outside of total visual accessibility) is to look at running lines in relation to class levels,with every up and down tendency of form cycle you know throw in as a possibility,to actually get a genuine feel for the personality of the horse,what the connections infer about him or her based on their movement,and only then can a class assessment be made about the horse.Then all the objeective work be done in relating that to the field.As you and others have mentioned, over time scanning PP's bottom to top to get this feel is not burdensome and tedious.For me it instills the feeling of confidence in my assessment of horses.It determines the big or little of my bets.And the further we get into the computer age the bigger the overlay it becomes as past performances are proudly thrown out as yesterday's tools.It amazes me at times.I look at horses ,uch the same way as I look at boxers.There is no technological advancement that renders useless a jab,left hook,right hand,or in boxer's terms for class=ring generalship and the ability to take a punch.As much as I have become specific in my playing forte my approach is always meant to be universally applicable,and mainly because I like to play different tracks on a regular basis.Taking things like class assessment and looking at running lines and an array of pace figures(not compound rankings),these things make me feel as close as humanly possible to seeing into the raw reality of this game.Sorry for the long winded post,but this conversation stimulates me in so many ways.Awesome.

There is an advantage in such study that many may not be aware of, especially if you are looking at races from a favorite track. It doesn't take long to realize that what superficially looks like a "move in class" is really not, or that a race in what superficially looks like another race in the same class is really a move up or down.

When I was playing Tup steadily (and seriously) I found six distinct levels of class in the "bottom" claiming layer, that were often more difficult to move up in than it was to jump from that level to the next "recognized" level. Most people just glance at the claiming price and label it as whatever, not realizing that variations in race conditions can create very specific divisions within (what seems on the surface to be) the same class.

Any effort expended in that direction can only increase one's knowledge of racing in general, and class levels in particular. Combining that with an understanding of why a trainer enters a specific horse in a specific race at a specific time makes racing a lot more sensible (and a lot more predictable).

[traynor]

I believe that a numerical rating can be given for class as well as pace/speed. Because I make one and it works for me, especially in route and turf races.

When I handicap I use a variation of the Performance Class Rating (PRC) from William L. Scott's "Total Victory at the track.)

To make the task simpler I created a comparison chart of class levels, though nothing is set in stone.

Class levels are in 20% increments higher and lower.

A $6,250 Stakes race is equivalent to a $8k Alw and a $10k starter Alw or OpClm and $16k open claiming and a $25k Md Spc and a $80k mdn Clm.

My chart runs from 1==Grade 1 races to 30 $1k claiming and $3200 Mdn clm.

I adjust one level down for NW3L, and two levels down for state bred and NW2L. Example a 50k state bred NW2L would be equivalent to an open $20k claiming race or a 32k mdn special race.

And I might adjust for track to track. A track with 25k claimers running for $30k as against a $25 claimer running at a different track for $13k might have an adjustment.

Then I make my pace/class figures and if a horse is one of the top horses in both categories and is in form, then I have a strong contender.

I don't have rules when I handicap. I have guidelines.

One of Joseph Finley's big contributions to racing was the emphasis he placed on five furlong workouts--especially in under a minute. I cashed more than a few tickets on first or second time starters that had posted a sub-minute 5f work a day or two before the race. Life seemed much easier then.

[JohnGalt1]

What I do is see if a maiden has the fastest 5f wo (by at least 2/5th) within the last 30 days. Unless the fastest is really the least slowest.

I bet $40 on a first timer Sunday at Presque Isle Downs in the third race, the #6 horse Ft Deposit. On June 6 it ran a 59.1 B work out. It paid $13.20.

It still works.

[Capper Al]

I believe that a numerical rating can be given for class as well as pace/speed. Because I make one and it works for me, especially in route and turf races.

When I handicap I use a variation of the Performance Class Rating (PRC) from William L. Scott's "Total Victory at the track.)

To make the task simpler I created a comparison chart of class levels, though nothing is set in stone.

Class levels are in 20% increments higher and lower.

A $6,250 Stakes race is equivalent to a $8k Alw and a $10k starter Alw or OpClm and $16k open claiming and a $25k Md Spc and a $80k mdn Clm.

My chart runs from 1==Grade 1 races to 30 $1k claiming and $3200 Mdn clm.

I adjust one level down for NW3L, and two levels down for state bred and NW2L. Example a 50k state bred NW2L would be equivalent to an open $20k claiming race or a 32k mdn special race.

And I might adjust for track to track. A track with 25k claimers running for $30k as against a $25 claimer running at a different track for $13k might have an adjustment.

Then I make my pace/class figures and if a horse is one of the top horses in both categories and is in form, then I have a strong contender.

I don't have rules when I handicap. I have guidelines.

Good stuff.

[pandy]

I found this thread interesting. I'm in the process of writing a book on how to use class in handicapping. The obvious is, many winners are horses that have a class edge and that class edge is simple: in the not-to-distant past the horse was facing better fields. This is a truth of handicapping and the reason why trainers drop their horses in class, they want to win.

The key is deciding whether the horse was competitive enough against those tougher fields to warrant a "class edge" today.

Pace has a lot to do with it. As the class level gets tougher the pace pressure gets stronger. Classy horses deal with pace pressure better.

In most of the medium to higher classes, it's how the horse finishes against how fast it runs early. This is the key. Horses with class edges win because they finish better than the other horses (in relation to the pace of the race).

Seattle Slew was not a fast finisher. True or false? It can be answered both ways. In simple terms, he was not a fast finisher. But he was a fast finisher against the pace of the race. In other words, he was the only horse that could finish as well as he did after running as fast as he did pre-stretch. Consequently, he was near unbeatable. It's always the same. Name any top-class horse, the great ones that rarely lost. Slew, Secretariat, Ruffian, Dr. Fager, Zenyatta, Citation, and all great horses, they all have one thing in common. They finished faster against the pace of the race than the other horses. And this can be quantified.

Thoroughbred or harness racing, same thing. If you look at Captain Courageous win the other night in the North American Cup. The last 1/16 of that race was slow, the pace just collapsed. Captain Courageous did not actually finish that well, he was all-out and basically horses just came back to him. But, when you compare his stretch kick to his pre-stretch energy output, he finished better against the pace than any horse in the race, which is why he won. Wake Up Peter actually gained more ground in the stretch, paced a faster last quarter by 2/5's, and probably raced a longer mile. But he was father off the pace so he had less pre-stretch energy expenditure. So Captain Courageous did actually finish better against the pace of the race and against his own pace figures, which is why he won and why he now has 11 wins in 13 career starts.

[traynor]

What I do is see if a maiden has the fastest 5f wo (by at least 2/5th) within the last 30 days. Unless the fastest is really the least slowest.

I bet $40 on a first timer Sunday at Presque Isle Downs in the third race, the #6 horse Ft Deposit. On June 6 it ran a 59.1 B work out. It paid $13.20.

It still works.

It was easier when fewer people looked for it. A bullet 5f is like a huge neon sign with a loudspeaker blaring, "Look at me!" I'll take any 5f under a minute, or a 4f under 48, if I know the trainer is (relatively) honest. It is still a very strong selection criteria for dirt sprints in the desert southwest, not so much SoCal. There are a number of trainers (who seem able to be totally ignored) on the southwest circuit who routinely use a 5f workout as prep for an upcoming 6 or 6.5, usually no more than a few days prior. Not quite as good as printing one's own money, but pretty close. Finding a few such trainers can go a long way toward improving one's bottom line.

[Capper Al]

I found this thread interesting. I'm in the process of writing a book on how to use class in handicapping. The obvious is, many winners are horses that have a class edge and that class edge is simple: in the not-to-distant past the horse was facing better fields. This is a truth of handicapping and the reason why trainers drop their horses in class, they want to win.

The key is deciding whether the horse was competitive enough against those tougher fields to warrant a "class edge" today.

Pace has a lot to do with it. As the class level gets tougher the pace pressure gets stronger. Classy horses deal with pace pressure better.

In most of the medium to higher classes, it's how the horse finishes against how fast it runs early. This is the key. Horses with class edges win because they finish better than the other horses (in relation to the pace of the race).

Seattle Slew was not a fast finisher. True or false? It can be answered both ways. In simple terms, he was not a fast finisher. But he was a fast finisher against the pace of the race. In other words, he was the only horse that could finish as well as he did after running as fast as he did pre-stretch. Consequently, he was near unbeatable. It's always the same. Name any top-class horse, the great ones that rarely lost. Slew, Secretariat, Ruffian, Dr. Fager, Zenyatta, Citation, and all great horses, they all have one thing in common. They finished faster against the pace of the race than the other horses. And this can be quantified.

Thoroughbred or harness racing, same thing. If you look at Captain Courageous win the other night in the North American Cup. The last 1/16 of that race was slow, the pace just collapsed. Captain Courageous did not actually finish that well, he was all-out and basically horses just came back to him. But, when you compare his stretch kick to his pre-stretch energy output, he finished better against the pace than any horse in the race, which is why he won. Wake Up Peter actually gained more ground in the stretch, paced a faster last quarter by 2/5's, and probably raced a longer mile. But he was father off the pace so he had less pre-stretch energy expenditure. So Captain Courageous did actually finish better against the pace of the race and against his own pace figures, which is why he won and why he now has 11 wins in 13 career starts.

Class is based on performance within competitive fields. If figured correctly, it's not the horse that is getting evaluated but rather the horse's relation to the field. What's great about being able to have the winner in my top two picks 50% of time for non-maiden races is not the hit or strike ratio. It is its independence from time based calculations.

[Augenj]

Class is based on performance within competitive fields. If figured correctly, it's not the horse that is getting evaluated but rather the horse's relation to the field. What's great about being able to have the winner in my top two picks 50% of time for non-maiden races is not the hit or strike ratio. It is its independence from time based calculations.
Agreed. :ThmbUp:

[traynor]

I found this thread interesting. I'm in the process of writing a book on how to use class in handicapping. The obvious is, many winners are horses that have a class edge and that class edge is simple: in the not-to-distant past the horse was facing better fields. This is a truth of handicapping and the reason why trainers drop their horses in class, they want to win.

The key is deciding whether the horse was competitive enough against those tougher fields to warrant a "class edge" today.

Pace has a lot to do with it. As the class level gets tougher the pace pressure gets stronger. Classy horses deal with pace pressure better.

In most of the medium to higher classes, it's how the horse finishes against how fast it runs early. This is the key. Horses with class edges win because they finish better than the other horses (in relation to the pace of the race).

Seattle Slew was not a fast finisher. True or false? It can be answered both ways. In simple terms, he was not a fast finisher. But he was a fast finisher against the pace of the race. In other words, he was the only horse that could finish as well as he did after running as fast as he did pre-stretch. Consequently, he was near unbeatable. It's always the same. Name any top-class horse, the great ones that rarely lost. Slew, Secretariat, Ruffian, Dr. Fager, Zenyatta, Citation, and all great horses, they all have one thing in common. They finished faster against the pace of the race than the other horses. And this can be quantified.

Thoroughbred or harness racing, same thing. If you look at Captain Courageous win the other night in the North American Cup. The last 1/16 of that race was slow, the pace just collapsed. Captain Courageous did not actually finish that well, he was all-out and basically horses just came back to him. But, when you compare his stretch kick to his pre-stretch energy output, he finished better against the pace than any horse in the race, which is why he won. Wake Up Peter actually gained more ground in the stretch, paced a faster last quarter by 2/5's, and probably raced a longer mile. But he was father off the pace so he had less pre-stretch energy expenditure. So Captain Courageous did actually finish better against the pace of the race and against his own pace figures, which is why he won and why he now has 11 wins in 13 career starts.

There are a number of misconceptions in the fundamental principles in pace analysis that make pace (as ordinarily calculated and understood) of questionable value in establishing class as a derivative.

[pandy]

Pace handicappers tend to only look at early pace. Late pace in relation to early pace is a good way to analyze both form and class.

[speed]

How do i determine class.
I walk into a bar and the women who ignore me, we'll, they have class. :)

[speed]

That's why when there is a filly race i go out to the walking ring and the horses who ignore me, i mean the one's that won't even look at me. I box. :lol:

[traynor]

Pace handicappers tend to only look at early pace. Late pace in relation to early pace is a good way to analyze both form and class.

On the contrary, pace handicappers who wager on their pace analysis tend to obsessively wring every bit of information they can from the available data, and have been doing so for many years.

[so.cal.fan]

I attempt to determine class as the horse's ability to carry it's speed over the distance of ground nature intended it to.
If you can classify horses correctly, you can win money at the track.

[classhandicapper]

If it was easy, every kid on the block would be doing it, and it would have no value. The further I go, the more I realize how much an advantage it is that a lot of handicappers are stuck back in the old days of speed numbers, simplistic pace analysis, and recycled strategies that have not changed much since the days of Ainslie, Beyer, Quirin, Sartin, et al. Or making great discoveries based on samples of a few hundred races.

It makes the work more rewarding for those willing to do it.

The irony is that most numbers based handicappers think of class handicappers as dinosaurs and fools, but that's primarily because very little has been written on the subject of class that was any good in a long time.

There have been an endless number of books written advancing the art of making pace numbers, final time numbers, velocity rating etc... You can buy good figures from a variety of sources. But it's not like the class handicappers are still dividing earning by the number of starts like they were in the 60s. They have advanced just as much. It just requires a lot more work to be done well and get a solid edge.

[JackS]

For me it's easier to consider ALL horses as claimers and to assume that 90% of all allowance runners will be running claiming as soon as they prove themselves unworthy of anything beyond NW1.
Most horses will show who they are and what they may be eventually worth wiith their first maiden win. I peg them at $25k at the majors and these same horses at $32k NW1, $50k NW2. The caveat with horses who look like this and all of a sudden are entered for claiming, is trainer intentions. It would be reasonable to expect all Nw's to continue to advance in non-claiming races.
These same tenents extend into Graded stakes particularly G2/3 which often show up in the C80-100k races. They are running for these claiming prices because thats all they're worth and in many cases worth much less possibly because of age or developing/developed infirmities .
I suggest that everyone look at all horses as claimers and try your best to put a price on them even if they have never been entered for a claiming price.

[classhandicapper]

I found this thread interesting. I'm in the process of writing a book on how to use class in handicapping. The obvious is, many winners are horses that have a class edge and that class edge is simple: in the not-to-distant past the horse was facing better fields. This is a truth of handicapping and the reason why trainers drop their horses in class, they want to win.

The key is deciding whether the horse was competitive enough against those tougher fields to warrant a "class edge" today.

Pace has a lot to do with it. As the class level gets tougher the pace pressure gets stronger. Classy horses deal with pace pressure better.

In most of the medium to higher classes, it's how the horse finishes against how fast it runs early. This is the key. Horses with class edges win because they finish better than the other horses (in relation to the pace of the race).

Seattle Slew was not a fast finisher. True or false? It can be answered both ways. In simple terms, he was not a fast finisher. But he was a fast finisher against the pace of the race. In other words, he was the only horse that could finish as well as he did after running as fast as he did pre-stretch. Consequently, he was near unbeatable. It's always the same. Name any top-class horse, the great ones that rarely lost. Slew, Secretariat, Ruffian, Dr. Fager, Zenyatta, Citation, and all great horses, they all have one thing in common. They finished faster against the pace of the race than the other horses. And this can be quantified.

Thoroughbred or harness racing, same thing. If you look at Captain Courageous win the other night in the North American Cup. The last 1/16 of that race was slow, the pace just collapsed. Captain Courageous did not actually finish that well, he was all-out and basically horses just came back to him. But, when you compare his stretch kick to his pre-stretch energy output, he finished better against the pace than any horse in the race, which is why he won. Wake Up Peter actually gained more ground in the stretch, paced a faster last quarter by 2/5's, and probably raced a longer mile. But he was father off the pace so he had less pre-stretch energy expenditure. So Captain Courageous did actually finish better against the pace of the race and against his own pace figures, which is why he won and why he now has 11 wins in 13 career starts.

Spoken like a guy that has spent a lot of time studying and betting harness racing (pre stretch energy used vs. how the finished). I mean that as a compliment. :lol:

I agree wholeheartedly with this. My only addition is that sometimes it's hard to measure all this stuff at the horse level by looking at "times". Sometimes it's easier and better to just compare the horses' trips and finishes.

[raybo]

On the contrary, pace handicappers who wager on their pace analysis tend to obsessively wring every bit of information they can from the available data, and have been doing so for many years.

Oh Traynor, you know that we just look at the 2nd call and head for the windows. Geez doesn't everybody do it that way?

[raybo]

The irony is that most numbers based handicappers think of class handicappers as dinosaurs and fools, but that's primarily because very little has been written on the subject of class that was any good in a long time.

There have been an endless number of books written advancing the art of making pace numbers, final time numbers, velocity rating etc... You can buy good figures from a variety of sources. But it's not like the class handicappers are still dividing earning by the number of starts like they were in the 60s. They have advanced just as much. It just requires a lot more work to be done well and get a solid edge.

Sorry you feel that way, but maybe some of us think class is best expressed as a function of other methods. And, some of us think that class can be expressed numerically. If there are not books being written specifically about class, doesn't mean that class is not covered very well in other ways. I think that class, in all it's forms, is ingrained in almost every decent handicapping book written, they may not call it "class", but it's still there.

[thaskalos]

On the contrary, pace handicappers who wager on their pace analysis tend to obsessively wring every bit of information they can from the available data, and have been doing so for many years.

The amusing thing about handicapping is that every single factor of the process can be as simple, or as complicated, as the practitioner wants to make it.

You got players who only look at the Beyer figure of the horse's last "qualified" race...and yet they consider themselves to be "speed handicappers".

Others only look to see who figures to take the lead today...and they call this "pace handicapping".

And when they make the determination that closers should be favored when a race is loaded with early speed...or that the lone front runner should be played in an otherwise "paceless" field...they think that they are engaging in "advanced" pace handicapping theory.

And when they see that their simplistic methods fail to deliver the desired results...then these players cry out that speed and pace handicapping have lost their bite...because "everybody is doing it"...

Funny stuff...

[raybo]

The amusing thing about handicapping is that every single factor of the process can be as simple, or as complicated, as the practitioner wants to make it.

You got players who only look at the Beyer figure of the horse's last "qualified" race...and yet they consider themselves to be "speed handicappers".

Others only look to see who figures to take the lead today...and they call this "pace handicapping".

And when they make the determination that closers should be favored when a race is loaded with early speed...or that the lone front runner should be played in a paceless field...they think that they are engaging in "advanced" pace handicapping theory.

And when they see that their simplistic methods fail to deliver the desired results...then these players cry out that speed and pace handicapping have lost their bite...because "everybody is doing it"...

Funny stuff...

Agree! :ThmbUp:

[classhandicapper]

Sorry you feel that way, but maybe some of us think class is best expressed as a function of other methods. And, some of us think that class can be expressed numerically. If there are not books being written specifically about class, doesn't mean that class is not covered very well in other ways. I think that class, in all it's forms, is ingrained in almost every decent handicapping book written, they may not call it "class", but it's still there.

I wasn't clear.

There are many methods for measuring the ability/performances of horses, but most of the popular methods these days are based on the "times" of races (fractional and final times). You can call that "class" handicapping (and at some level it is), but that's not what it is in a traditional sense.

Traditionally, class handicapping was about the techniques that can be used for determining the quality of a field, how well a horse ran within the race, etc... in a NON TIME based way. Essentially looking at who beat who, by how much, how consistently, and with what trip.

Very little has been written about that approach in a comprehensive way in decades.

Essentially, decades ago most time based methods were woefully inaccurate. So many handicappers tried to evaluate field quality and performance using claiming prices, race conditions, purses, earnings, consistency record etc...

Then the time based methods exploded in sophistication, became way more popular and effective, and were easily available. So the conventional wisdom has been that the more traditional class handicapping method was outdated and inferior.

I am arguing that's it's not. The same advancements have also been occurring in class handicapping, but no one is writing about it in a comprehensive way because it's much harder to package class into a very high quality number that can be sold. It also requires WAY more time and energy to do well.

[LottaKash]

Very little has been written about that approach in a comprehensive way in decades.



Yes, and that is why I had to find my own way in regards of assessing "true-class"....Once over that hurdle, the game became much, much easier for me...But it took such a long time tho...Still, better late than never, I am pleased to say...:cool:

And, finding true class, in the end, turns out be a most simplistic affair for me when compared to the many theories that I have read and pondered about thru the years....

I just don't honestly believe that any computer algorithim, could do it any better than me.....Finding the True Class in any given situation, is not a "secret-hidden" thing that a computer will reveal imo...Class is, as Class does...

[rico]

Hello gents, just a few words,

If you can get your hands on Frederick S. Davis "Thoroughbred Racing:

Probability Computation" your problems will be a thing of the past.

[MJC922]

I am arguing that's it's not. The same advancements have also been occurring in class handicapping, but no one is writing about it in a comprehensive way because it's much harder to package class into a very high quality number that can be sold. It also requires WAY more time and energy to do well.

Well stated. I know it's probably to be scoffed at because many these days believe horse against horse stuff and pecking order is a crude way to get a handle on this -- and on one hand I don't disagree with that, it was too crude for my taste buds, but on the other hand it's at the very essence of what we're trying to quantify. I personally found that I had to use a computer to make class ratings with enough precision to rival the best speed figures. Unlike Picking Winners though, a book which literally put the keys into the hands of everyone, I suspect class ratings when done on a grand scale like this would not sell many books as it would take an army to calculate them manually.

[MJC922]

I found this thread interesting. I'm in the process of writing a book on how to use class in handicapping. The obvious is, many winners are horses that have a class edge and that class edge is simple: in the not-to-distant past the horse was facing better fields. This is a truth of handicapping and the reason why trainers drop their horses in class, they want to win.

The key is deciding whether the horse was competitive enough against those tougher fields to warrant a "class edge" today.

Pace has a lot to do with it. As the class level gets tougher the pace pressure gets stronger. Classy horses deal with pace pressure better.

In most of the medium to higher classes, it's how the horse finishes against how fast it runs early. This is the key. Horses with class edges win because they finish better than the other horses (in relation to the pace of the race).

Seattle Slew was not a fast finisher. True or false? It can be answered both ways. In simple terms, he was not a fast finisher. But he was a fast finisher against the pace of the race. In other words, he was the only horse that could finish as well as he did after running as fast as he did pre-stretch. Consequently, he was near unbeatable. It's always the same. Name any top-class horse, the great ones that rarely lost. Slew, Secretariat, Ruffian, Dr. Fager, Zenyatta, Citation, and all great horses, they all have one thing in common. They finished faster against the pace of the race than the other horses. And this can be quantified.

Thoroughbred or harness racing, same thing. If you look at Captain Courageous win the other night in the North American Cup. The last 1/16 of that race was slow, the pace just collapsed. Captain Courageous did not actually finish that well, he was all-out and basically horses just came back to him. But, when you compare his stretch kick to his pre-stretch energy output, he finished better against the pace than any horse in the race, which is why he won. Wake Up Peter actually gained more ground in the stretch, paced a faster last quarter by 2/5's, and probably raced a longer mile. But he was father off the pace so he had less pre-stretch energy expenditure. So Captain Courageous did actually finish better against the pace of the race and against his own pace figures, which is why he won and why he now has 11 wins in 13 career starts.

Enjoyable post. I like where you stated the not-too-distant past line at the top. Pittsburg Phil said class and form are the twin sisters of handicapping. That would suggest class is only half the story in ranking the field. I'm of the opinion that even if God circled the program number of the class of the field for us and we didn't have to question it or lift a finger it's likely a third of the time the horse is in sub-par form and runs flat, a third of the time the horse is in moderate form but still gets beat by some other horse of slightly less class which throws a peak effort and then finally you get your last third when the horse gets the job done. Probably a break even proposition. We need to know who is very likely outclassed, but after that give me sharp form any day.

[pondman]

And when they see that their simplistic methods fail to deliver the desired results...then these players cry out that speed and pace handicapping have lost their bite...because "everybody is doing it"...

Funny stuff...
There is a reason for keeping class independent of all methods. Pick whatever relationship you want:

Portland to Grants
Hollywood to Golden Gate
Belmont to Delaware
Belmont to Churchill

They all have something in common The class horse, which lost by 20 lengths, can toy with the field and run from anywhere, and can be ridden by anyone who can hang on. It's a big mistiake to consider anything else. When I'm on a horse and someone tells me it's not fast enough, it tells me the individual doesn't understand class. Because class can be play entirely independent of a performance. The horse is in a soft spot.

[Magister Ludi]

The performance of thoroughbred racehorses (“TBR”) can be decomposed into three broad categories: ability, pecuniary incentives, and competition as a result of entropy. Ability accounts for about 67% of a TBR’s performance. It includes such factors as age, gender, weight, etc. Pecuniary incentives account for about 25% of a TBR’s performance. Horses can be managed by jockeys in a way that optimizes marginal monetary costs and benefits. Competition as a result of entropy comprises about 8% of a TBR’s performance.

There are several ways to measure the entropy of a race. It can be measured ex ante by the variance in the odds of the entrants. It can be measured ex post by the variance of performance at the finish.

As you well know, some TBR’s rise to the occasion while others wither in the heat of battle. Perhaps this is an important component of a TBR’s innate class.

[LottaKash]

There are several ways to measure the entropy of a race. It can be measured ex ante by the variance in the odds of the entrants. It can be measured ex post by the variance of performance at the finish.

.

Magister, I just came back from an aftertnoon birthday party, and I don't usually drink much anymore, but today I did...So please, in low level English, could you please explain this cryptic, to me as of now, message..Please...:jump:

[VeryOldMan]

The performance of thoroughbred racehorses (“TBR”) can be decomposed into three broad categories: ability, pecuniary incentives, and competition as a result of entropy. Ability accounts for about 67% of a TBR’s performance. It includes such factors as age, gender, weight, etc. Pecuniary incentives account for about 25% of a TBR’s performance. Horses can be managed by jockeys in a way that optimizes marginal monetary costs and benefits. Competition as a result of entropy comprises about 8% of a TBR’s performance.

There are several ways to measure the entropy of a race. It can be measured ex ante by the variance in the odds of the entrants. It can be measured ex post by the variance of performance at the finish.

As you well know, some TBR’s rise to the occasion while others wither in the heat of battle. Perhaps this is an important component of a TBR’s innate class.

I've seen enough of your posts to respect your analysis - could you please explain the "competition as a result of entropy" factor in plainer terms? I think I understand the "pecuniary incentives" point to some extent, but am struggling to understand the entropy point. Randomness/unknown?

[Magister Ludi]

>LottaKash: ex ante = before the fact
ex post = after the fact

>VeryOldMan: Entropy is disorder in a closed system. Think of it as a measure of the competitiveness of a race. A maximum entropy race would be a race where all of the horses have the same probability of winning. A minimum entropy race would be a uniform distribution of probabilities over the interval 0 to 1.

One way to quantify the competitivess of a race is with the following formula, presented by Dr. Beav, who used to post on this forum:

Sum{1/[O(i)+1]^2}/N

where O(i) = odds of the ith horse
N = number of entries

The smaller the value, the more competitive the race.

[traynor]

>LottaKash: ex ante = before the fact
ex post = after the fact

>VeryOldMan: Entropy is disorder in a closed system. Think of it as a measure of the competitiveness of a race. A maximum entropy race would be a race where all of the horses have the same probability of winning. A minimum entropy race would be a uniform distribution of probabilities over the interval 0 to 1.

One way to quantify the competitivess of a race is with the following formula, presented by Dr. Beav, who used to post on this forum:

Sum{1/[O(i)+1]^2}/N

where O(i) = odds of the ith horse
N = number of entries

The smaller the value, the more competitive the race.

I think that when the day comes that I use the tote board (and that profound and mysterious phenomena referred to as the foolishness of crowds) as the criteria for quantifying the competitiveness of a race, I will give it up and go back to playing blackjack full-time.

[LottaKash]

[QUOTE=Magister Ludi

The smaller the value, the more competitive the race.[/QUOTE]

Magister, thx for the lowdown....You sure can dazzle a crowd with your footwork....:jump: ...

In essence, and sincerely, I understand now....but, why couldn't you explain that in in layman's terms....Still, I know of what you speak, because those things have crossed my mind, a time or two......

[raybo]

Magister, thx for the lowdown....You sure can dazzle a crowd with your footwork....:jump: ...

In essence, and sincerely, I understand now....but, why couldn't you explain that in in layman's terms....Still, I know of what you speak, because those things have crossed my mind, a time or two......

That's why some brilliant minds can teach, and other brilliant minds can't. Some of them can't relate terms in commonly understandable language, while some of them can.

I have a brother-in-law (EE) who worked for NASA at one time, who cannot teach me s--t, because he does not have the ability to relate highly technical terms in commonly understandable language. In other words, he talks "over my head".

[thaskalos]

I think that when the day comes that I use the tote board (and that profound and mysterious phenomena referred to as the foolishness of crowds) as the criteria for quantifying the competitiveness of a race, I will give it up and go back to playing blackjack full-time.

That day might be closer than you think... :)

[BlueChip@DRF]

Was last night's 8th race at Mountaineer an example of "back class"?

[traynor]

That day might be closer than you think... :)

I really hope you are wrong. For all its many faults, shortcomings, and deficiencies, it is still way easier to make money wagering on races than playing blackjack.

[Robert Fischer]

I actually use odds as a criteria(as opposed to the criterion) when quantifying the competitiveness of a race(post race). This is a process I currently do manually. It is accurate to the degree I require, and the fastest method.

I have thought about different options if I was ever to sit down with someone and automate the process. A power ranking for each individual horse would be the obvious substitution for the tote board. Actual race performance measures, and tote odds could also play a role in more sophisticated programming.

[PICSIX]

This one is easy IMO

[PICSIX]

This one is easy IMO

How did you determine he was the class horse? Better yet, how did he go off at over 5-1?

[traynor]

I actually use odds as a criteria(as opposed to the criterion) when quantifying the competitiveness of a race(post race). This is a process I currently do manually. It is accurate to the degree I require, and the fastest method.

I have thought about different options if I was ever to sit down with someone and automate the process. A power ranking for each individual horse would be the obvious substitution for the tote board. Actual race performance measures, and tote odds could also play a role in more sophisticated programming.

Based on results over an extended period of time, the greater the degree of separation between the performance indicators and the tote board, the higher the degree of accuracy. Mixing the use of objective performance indicators with the subjective opinions of a group that manages to lose (95%? 98%?) overall does not seem rational, much less useful.

There are no simple solutions (that I know about). Creating projected performance expectations for a group of horses in a particular race is slow, tedious, painstaking work. The upside is that the same slow, tedious, painstaking work can also generate "pace figures" and "performance envelopes" that make most "pace analysis" seem almost comically inept in comparison. Lots of hard work, but the rewards are commensurate with the effort invested.

One has to have a profound respect for those using antiquated and simplistic forms of "pace analysis" or "class analysis." Think of them as the cannon fodder bringing their money to the pari-mutuel battlefield. Even if they are a scruffy and smelly lot, we need them. They provide the profits that serious bettors remove from that battlefield every racing day.

[thaskalos]

I really hope you are wrong. For all its many faults, shortcomings, and deficiencies, it is still way easier to make money wagering on races than playing blackjack.
How is your poker game these days?

That might be our only reasonable choice left...

[raybo]

How did you determine he was the class horse? Better yet, how did he go off at over 5-1?

Well, although I am not crazy about man-made class levels, an open OC32K is way above Clm6250n1y. He has never run this low, in the races posted anyway (although I am not knowledgeable of the Brazilian race levels). His odds were higher probably because of the suspicious class drop and the fact that he finished last, or next to last, in his last 3 races in the US.

2nd ranked Prime Power, with poor performances, is a key to class also.

[raybo]

Based on results over an extended period of time, the greater the degree of separation between the performance indicators and the tote board, the higher the degree of accuracy. Mixing the use of objective performance indicators with the subjective opinions of a group that manages to lose (95%? 98%?) overall does not seem rational, much less useful.

There are no simple solutions (that I know about). Creating projected performance expectations for a group of horses in a particular race is slow, tedious, painstaking work. The upside is that the same slow, tedious, painstaking work can also generate "pace figures" and "performance envelopes" that make most "pace analysis" seem almost comically inept in comparison. Lots of hard work, but the rewards are commensurate with the effort invested.

One has to have a profound respect for those using antiquated and simplistic forms of "pace analysis" or "class analysis." Think of them as the cannon fodder bringing their money to the pari-mutuel battlefield. Even if they are a scruffy and smelly lot, we need them. They provide the profits that serious bettors remove from that battlefield every racing day.

Although I don't completely understand what you meant in the first sentence, I agree with the rest. :ThmbUp:

[raybo]

How is your poker game these days?

That might be our only reasonable choice left...

So, are you saying that the game is becoming unbeatable? That's impossible isn't it, in a paramutuel system? There must be winners and losers in a paramutuel system, wouldn't you say? Or, are you just saying it is becoming harder and harder to remain profitable?

[DeltaLover]

So, are you saying that the game is becoming unbeatable? That's impossible isn't it, in a paramutuel system? There must be winners and losers in a paramutuel system, wouldn't you say? Or, are you just saying it is becoming harder and harder to remain profitable?

It is quite possible.

Only one winner must be, which is the take out. More than this, if the crowd reaches a certain level of accuracy, the game becomes unbeatable.

Hopefully, T-Bred betting is still far from this...

[thaskalos]

So, are you saying that the game is becoming unbeatable? That's impossible isn't it, in a paramutuel system? There must be winners and losers in a paramutuel system, wouldn't you say? Or, are you just saying it is becoming harder and harder to remain profitable?
As DeltaLover already stated...the "house" is the only guaranteed winner in a gambling game -- whether this game is parimutuel or not.

The onerous takeout, the increased sophistication of the higher-betting crowd, and the dropping out of the unsophisticated players all conspire to make the profit margins in this game slimmer and slimmer as time passes.

I am an every day player...and I don't like what I see in this game during the weekdays. Short and uncompetitive fields are not exactly the places where the "sophisticated" player can apply his skill to outplay the crowd by the margin needed to overcome the takeout.

With the casino money being more widespread, I see the six-horse fields sprouting up everywhere...and even the unsophisticated player can make sense of fields of this size.

IMO...the "winning" horseplayers are standing on shaky ground...even if most haven't realized it yet.

[traynor]

How is your poker game these days?

That might be our only reasonable choice left...

I started betting on harness races because my employer at the time favored them, having won substantial amounts. My employer and his associates were all "gamblers" in the sense that they only bet to win. If they couldn't win honestly, they thought nothing at all of cheating. And they had developed their skills to high art. "Mechanics" does not do them justice. They were artists.

I watched them, I listened to their explanations of what they were doing, and it still took repeated slow motion demonstrations before I could actually (consciously) see what they were doing. I watched them fleece a steady stream of marks, amateur, professional, and every flavor in between.

I have seen a few decent mechanics in casinos, mostly the smaller ones, or those located in specific countries outside the US.

That is a long, convoluted way of saying I don't play private games, so that makes casinos almost the only option. Some legal card rooms in some venues exist, but anything over trivial amounts tends to be a small pool filled with large sharks. And more than a few barracuda.

I was fortunate to carry my paranoia about cards over to blackjack. I prospered by specializing in six- and eight-deck shoes, while all the rookies were chasing hand-held games--and losing steadily.

Poker is a good game. I just don't trust the other players. And when the other players can be trusted, there is always the guy with the sawyer who says, "Oh, look, Matilda! These nice folks gathered all the money together in one place for us. Isn't that nice?"

I much prefer baccarat, a game that "anyone who knows anything" declares "unbeatable" and will produce endless statistics to "prove" their point.

[traynor]

As DeltaLover already stated...the "house" is the only guaranteed winner in a gambling game -- whether this game is parimutuel or not.

The onerous takeout, the increased sophistication of the higher-betting crowd, and the dropping out of the unsophisticated players all conspire to make the profit margins in this game slimmer and slimmer as time passes.

I am an every day player...and I don't like what I see in this game during the weekdays. Short and uncompetitive fields are not exactly the places where the "sophisticated" player can apply his skill to outplay the crowd by the margin needed to overcome the takeout.

With the casino money being more widespread, I see the six-horse fields sprouting up everywhere...and even the unsophisticated player can make sense of fields of this size.

IMO...the "winning" horseplayers are standing on shaky ground...even if most haven't realized it yet.

Being an everyday player may be a part of the problem. Consider it the equivalent of hunting for food. Better to pass up the occasional shot at a rabbit or squirrel to have a shot at a deer, elk, or moose. The difficulty in hunting rabbits and squirrels is that there is never much chance to sit back with a full belly and doze off with dreams of long-forgotten battles in long-forgotten wars. One has to be out there all day, every day, chasing the bleeping little critters just to have something for dinner.

[thaskalos]

Being an everyday player may be a part of the problem. Consider it the equivalent of hunting for food. Better to pass up the occasional shot at a rabbit or squirrel to have a shot at a deer, elk, or moose. The difficulty in hunting rabbits and squirrels is that there is never much chance to sit back with a full belly and doze off with dreams of long-forgotten battles in long-forgotten wars. One has to be out there all day, every day, chasing the bleeping little critters just to have something for dinner.
It's a shame that the racing industry is totally unaware of the devastating affect this takeout has when it's apllied to today's five, six, and seven-horse fields.

When a poker game gets short-handed at a casino...the players ask for a rake reduction...and they get it.

[OCF]

It's a shame that the racing industry is totally unaware of the devastating affect this takeout has when it's apllied to today's five, six, and seven-horse fields.

When a poker game gets short-handed at a casino...the players ask for a rake reduction...and they get it.

Hmmm, I know next to zero about poker but it sounds like the takeout has some variability. I wonder if that would work for racing. Just thinking out loud.

[traynor]

It's a shame that the racing industry is totally unaware of the devastating affect this takeout has when it's apllied to today's five, six, and seven-horse fields.

When a poker game gets short-handed at a casino...the players ask for a rake reduction...and they get it.

I stopped betting California tracks years ago for just that reason--too many short fields, and pitiful exactas. I have never regretted it.

It is a similar situation in many other fields. As long as consumers are willing to buy shoddy merchandise, it is guaranteed that someone will produce (or continue producing) that shoddy merchandise. Complaints are useless, because the reality (people keep buying) makes the complaints little more than annoying. The only way to change the situation is directly--for everyone to stop betting completely on races with short fields.

Is that gong to happen? Perhaps in G.H.W. Bush's "kinder, gentler world." Not in this one.

[VeryOldMan]

As DeltaLover already stated...the "house" is the only guaranteed winner in a gambling game -- whether this game is parimutuel or not.

The onerous takeout, the increased sophistication of the higher-betting crowd, and the dropping out of the unsophisticated players all conspire to make the profit margins in this game slimmer and slimmer as time passes.

I am an every day player...and I don't like what I see in this game during the weekdays. Short and uncompetitive fields are not exactly the places where the "sophisticated" player can apply his skill to outplay the crowd by the margin needed to overcome the takeout.

With the casino money being more widespread, I see the six-horse fields sprouting up everywhere...and even the unsophisticated player can make sense of fields of this size.

IMO...the "winning" horseplayers are standing on shaky ground...even if most haven't realized it yet.

Great post.

We're getting far afield from the original topic here, but thanks for confirming what I thought I was seeing - really short fields at many tracks.

Don't we also have to consider one of the other groups of players at the table - the owners. I don't know how any owner can make money at a place like Beulah Park given the purses, but if I owned a horse there (for example) I'd be pretty happy to have my horse in a field no larger than 6 and hopefully shorter. May not be the best for the long-term health of the sport at some macro level, but I have to believe a lot of owners are happy to cash any check they can get and aren't clamoring for bigger fields. We're on the other side of the table from them.

[traynor]

Great post.

We're getting far afield from the original topic here, but thanks for confirming what I thought I was seeing - really short fields at many tracks.

Don't we also have to consider one of the other groups of players at the table - the owners. I don't know how any owner can make money at a place like Beulah Park given the purses, but if I owned a horse there (for example) I'd be pretty happy to have my horse in a field no larger than 6 and hopefully shorter. May not be the best for the long-term health of the sport at some macro level, but I have to believe a lot of owners are happy to cash any check they can get and aren't clamoring for bigger fields. We're on the other side of the table from them.

Your post is not off-topic. Many consider races with short fields to be less competitive, and therefore of lesser "class." As Quirin (and others later) pointed out, a horse that is fourth in a ten horse field has passed six horses--a horse that is fourth in a five horse field has only managed to run less slow than one other horse in the race. Field size--in all its various manifestations--is an intrinsic component of any meaningful attempt to determine race class. It is far more important than the subjective opinons of the crowd expressed on the tote board.

[raybo]

I can see the argument being true, in the future, if all bettors are undisciplined. But, there will always be some bettors who are disciplined, betting only when they are getting the value they require. The problem with this, of course, is fewer plays, and maybe smaller profits. But, there will always be a few winners in the game.

[Capper Al]

Any interest in posting class figs?

[MJC922]

Any interest in posting class figs?

Sure man, fire away. Are you thinking like the the top class rated efforts of the past x number of years? or stuff for tomorrow's card? The latter I can't do unfortunately, contract doesn't allow it.

[MJC922]

Attached my top 15 class rated efforts from 2008 until the present. Would like to see the top Beyers, Ragozin's etc over the same period if anyone has them and of course would like to see anyone else who makes their own ratings on the board.

[traynor]

Any interest in posting class figs?

I'm sure everyone would like to see what you have to offer. Please feel free to post any class ratings/class figures you have/use/have developed/are developing.

[raybo]

Any interest in posting class figs?

Sure, pick a race for tomorrow. Or, pick a whole card.

[Capper Al]

Understanding that one card won't be significant either way, if I do well or not, I'll post the top two for each of Saturday's AP non-maiden races. It would be nice if others joined in.

[traynor]

Understanding that one card won't be significant either way, if I do well or not, I'll post the top two for each of Saturday's AP non-maiden races. It would be nice if others joined in.

I think 50% winners in the top 2 is pretty significant, even in a small sample.

[Capper Al]

I think 50% winners in the top 2 is pretty significant, even in a small sample.

It should be.

[raybo]

Understanding that one card won't be significant either way, if I do well or not, I'll post the top two for each of Saturday's AP non-maiden races. It would be nice if others joined in.

There's a chance for rain Saturday in Chicago, so here's today's card for Belmont. Same understanding that Al stated, 1 card means very little in the long term, so, no guarantees, play these at your own risk! I will post the top 2 class rated horses (my own auto-calculated performance class ratings), unless there are virtual ties, in which case I will post more than 2.

I will say that these ratings include overall performance and class in competitive races, and an adjustment for form cycle and current form. There are no subjective judgements made by me, strictly computer generated.

BEL 06/20/2013

Race 1 8.50f InnerT SOC
1 142.8
8 138.0
7 137.7

Race 2 6.00f Dirt Clm 30000
2 126.8
8 126.0

Race 3 6.00f Dirt Clm 14000
5 141.3
7 135.3

Race 4 7.00f Turf SOC
3 281.4
5 153.9

Race 5 8.00f Dirt OClm 14000nw1
4 140.0
1 137.7

Race 6 8.50f InnerT Clm 20000
12 125.5
7 122.4

Race 7 8.00f Turf SOC
8 140.9
6 140.3

Race 8 8.00f Dirt OClm 25000nw2
7 157.8
10 153.7
4 153.3

Race 9 7.00f Turf Md Sp Wt
3 133.0
5 131.0
2 130.0

[raybo]

I think 50% winners in the top 2 is pretty significant, even in a small sample.

Why, all of a sudden, is it ok to accept a small sample, very small in this case, of being significant, when you are constantly raving about the unreliability of small samples? One single card is proof of nothing, except that you were right or wrong on that particular card.

Geez, make up your mind. :confused:

[Robert Fischer]

Even more important than the results for one card is whether the figs were correct/on the right path.

[raybo]

Even more important than the results for one card is whether the figs were correct/on the right path.

But, who or what determines what is correct or on the right path? Long term results right? Class, is like all other determinations, right sometimes and wrong other times. The important thing is, can the ratings be used to enhance profitability in the long term, but not necessarily on a single card, or in a single race.

Come on guys/gals, if you're going to state that a single card is proof of something, then post your own class ratings. Otherwise, you don't have much of a leg to stand on by saying a single card means anything or not. I'm sure, if you did post your ratings and your hit rate sucked on a single card, you'd probably say that it didn't mean the ratings were no good because it was just one card.

Why is class determination any different than the other ratings produced? No rating works the same on every card or on every race. If you disagree then you need to be presenting some kind of truth or reasoning behind your opinion.

The first one who made such a statement has made remarks in the past about incorrect or non-viable assumptions being made, by some, based on what he considers small samples, and yet here he states just the opposite, but has never presented any examples of his own work. Just seems like he thinks it's ok, for him, to come on here and make statements like they are fact, without concrete proof, but the same is not ok for anyone else here. A bit self-serving wouldn't you say.

I'm not trying to incite a riot in this thread, but let's get real, at least.

[thaskalos]

I think 50% winners in the top 2 is pretty significant, even in a small sample.

Wouldn't we have to look at the mutuel prices of these winners, before we conclude that this undertaking is "significant"?

I am pretty confident that I can pick 50% winners in the top 2, without any handicapping or ratings at all.

[Robert Fischer]

Why not, i'll give it a go

BEL 06/20/2013

Race 1 8.50f InnerT SOC
:9: 87
:7: 81

Race 2 6.00f Dirt Clm 30000
:9: 85
:7: 81

race 3 skip


Race 4 7.00f Turf SOC
:7: 90
:5: 85

Race 5
all you get is 3 races baby

[raybo]

Wouldn't we have to look at the mutuel prices of these winners, before we conclude that this undertaking is "significant"?

I am pretty confident that I can pick 50% winners in the top 2, without any handicapping or ratings at all.

That's my point exactly! If value means nothing, many here could pick 50% winners with 2 picks. However, if you're goal is to make money, hit rate means very little, especially on a single card of races. Heck, I could miss the winner in every race today, who knows? That wasn't my goal. My goal was to produce ratings that will make money long term. Hitting winners is the easy part, making long term profit is the hard part.

But, if everyone wants this to be about hitting winners, regardless of price, then that should have been stated beforehand. Anybody can come up with a class rating, but will it help you be profitable in the end? I didn't even consider the value in this card because we don't know what the odds will be yet, we'll see what happens I guess.

[Capper Al]

AP 21 Jun 13
...........................................*** .. e-Ponies
Race Cond..............cls I.......cls II ......cls...... ML

#1 Clm 7500n3L ....3-4-2 ....2-7-4 ....2-1-4 ...2-4-7
#2 Clm 7500n2l .....7-4-2 ....2-4-5 ....3-4-6 ...5-2-7
#3 OClm 80000b ....2-6-5 ....2-6-4 .... 2-5-6 ...2-3-6

#4 OClm 40000n2x .2-4-3 ....4-7-6 ... 1-[5-4]* 4-7-[5-3-6]*
#5 Clm 25000n2l ....3-4-5 ....4-5-3 ... 5-3-2 ....5-2-4

#6 Md 15000 .........5-1-4 .....4-1-6 ...1-3-4 ....3-4-5
#7 Md Sp Wt .........2-4-1..... 2-5-4 .. 9-2-8 ....4-5-8

#8 Clm 5000b .......1-9-7 .....7-1-5....1-7-2 ....[1-1a]*-7-4
#9 Clm 25000b .....10-3-5 ...4-9-3 ... 6-9-5 ....9-10-6

[]* is tied

'cls I' is kind of like EPS while 'cls II' should be better because it considers form a little. Also included are e-Ponies class picks and the Morning Line for comparison.

Note races #6 and #7 are maidens but included anyway for discussion. These numbers are good as long as a first time starter doesn't win. I'm working on a way to handle this when I get time.

Please remember these are not final picks only class ratings for discussion.

Good Luck!

[raybo]

Understanding that one card won't be significant either way, if I do well or not, I'll post the top two for each of Saturday's AP non-maiden races. It would be nice if others joined in.

Come on Al, you started this, post your ratings, or pick another track for today, doesn't matter to me. Why wait for Saturday, heck we can do that card too, on Saturday, even if it rains.

[raybo]

AP 21 Jun 13
...........................................*** .. e-Ponies
Race Cond..............cls I.......cls II ......cls...... ML

#1 Clm 7500n3L ....3-4-2 ....2-7-4 ....2-1-4 ...2-4-7
#2 Clm 7500n2l .....7-4-2 ....2-4-5 ....3-4-6 ...5-2-7
#3 OClm 80000b ....2-6-5 ....2-6-4 .... 2-5-6 ...2-3-6

#4 OClm 40000n2x .2-4-3 ....4-7-6 ... 1-[5-4]* 4-7-[5-3-6]*
#5 Clm 25000n2l ....3-4-5 ....4-5-3 ... 5-3-2 ....5-2-4

#6 Md 15000 .........5-1-4 .....4-1-6 ...1-3-4 ....3-4-5
#7 Md Sp Wt .........2-4-1..... 2-5-4 .. 9-2-8 ....4-5-8

#8 Clm 5000b .......1-9-7 .....7-1-5....1-7-2 ....[1-1a]*-7-4
#9 Clm 25000b .....10-3-5 ...4-9-3 ... 6-9-5 ....9-10-6

[]* is tied

'cls I' is kind of like EPS while 'cls II' should be better because it considers form a little. Also included are e-Ponies class picks and the Morning Line for comparison.

Note races #6 and #7 are maidens but included anyway for discussion. These numbers are good as long as a first time starter doesn't win. I'm working on a way to handle this when I get time.

Please remember these are not final picks only class ratings for discussion.

Good Luck!

So, which ratings would you be playing, if you were playing? Come on now, Robert and I went out on the limb, how about it? And, also, I don't have the 21st data files yet so can't post that card until then. But I will.

[Capper Al]

Wouldn't we have to look at the mutuel prices of these winners, before we conclude that this undertaking is "significant"?

I am pretty confident that I can pick 50% winners in the top 2, without any handicapping or ratings at all.

Can you do that without considering time(speed)? That's the point. What are your alternative points of view to time?

[Capper Al]

So, which ratings would you be playing, if you were playing? Come on now, Robert and I went out on the limb, how about it? And, also, I don't have the 21st data files yet so can't post that card until then. But I will.

I said 'cls II' is better, didn't I? e-Ponies and morning line are just added for comparision and discussion. The maiden races are outside of my stated claim. My claim is 50% top two in non-maiden races on class alone, no time/speed. I'll try to do Saturday's card also.

[Capper Al]

Why not, i'll give it a go

BEL 06/20/2013

Race 1 8.50f InnerT SOC
:9: 87
:7: 81

Race 2 6.00f Dirt Clm 30000
:9: 85
:7: 81

race 3 skip


Race 4 7.00f Turf SOC
:7: 90
:5: 85

Race 5
all you get is 3 races baby

Are these class ratings or picks?

[Capper Al]

My posted class ratings have not been adjusted by handicapping. They are straight out of my program as is. The reason a third horse was include is in case of scratches and, as any handicapper should know, you look closely at the horse that just missed being included.

Raybo and I agree on three races for top class pick: races 2, 5, and 8.

[thaskalos]

Can you do that without considering time(speed)? That's the point. What are your alternative points of view to time?

I can pick 50% winners with my top 2 selections without any handicapping at all. And I can do it for as long as you want...even over thousands of races.

And this includes every track in the country...and even Canada.

And furthermore...I can post my selections right here...in real time.

Guaranteed!

[Capper Al]

I can pick 50% winners with my top 2 selections without any handicapping at all. And I can do it for as long as you want...even over thousands of races.

And this includes every track in the country...and even Canada.

And furthermore...I can post my selections right here...in real time.

Guaranteed!

Join in. Post Friday's AP top two without using time (speed based), pace, fractional times, consensus picks, tote-board, or selector. Use just pure class.

[Light]

The top 2 betting favorites win 50% of the races statistically over the last several decades. Posting and winning 50% with your top 2 is snake oil.

[raybo]

Join in. Post Friday's AP top two without using time (speed based), pace, fractional times, consensus picks, tote-board, or selector. Use just pure class.

The point he was making, I assume, is picking 50% winners, not by class. I thought this thread was about class. Heck, the top 2 odds horses hit about 50% of the time don't they? That isn't the point of the thread, it's supposed to be about class, either traditional class methods or calculated class methods. At least that's what I thought anyway. Whatever!

[thaskalos]

The top 2 betting favorites win 50% of the races statistically over the last several decades. Posting and winning 50% with your top 2 is snake oil.

There you go! :ThmbUp:

[raybo]

I think everyone needs to consider that class determinations are only a portion of the equation to making money, and probably not even a large portion at that, for most players. Too often, the obvious class horses pay very little, and will lose you money long term. The point, for those who want to be profitable, is making money long term, regardless of the method used.

[raybo]

My posted class ratings have not been adjusted by handicapping. They are straight out of my program as is. The reason a third horse was include is in case of scratches and, as any handicapper should know, you look closely at the horse that just missed being included.

Raybo and I agree on three races for top class pick: races 2, 5, and 8.

My class horses are for Belmont today, yours are for AP tomorrow.

[thaskalos]

I think everyone needs to consider that class determinations are only a portion of the equation to making money, and probably not even a large portion at that, for most players. Too often, the obvious class horses pay very little, and will lose you money long term. The point, for those who want to be profitable, is making money long term, regardless of the method used.

The main problem, IMO, is that "class" is not the "constant" that many players consider it to be. You cannot "freeze-frame" a horse's class...because it is largely dependant on other handicapping factors as well. There is a lot of "overlapping" among the different handicapping factors...so you can not, in good conscience, use terms like "pure class"...or "pure form"...

[raybo]

The main problem, IMO, is that "class" is not the "constant" that many players consider it to be. You cannot "freeze-frame" a horse's class...because it is largely dependant on other handicapping factors as well. There is a lot of "overlapping" among the different handicapping factors...so you can not, in good conscience, use terms like "pure class"...or "pure form"...

I agree, that's why I made a form adjustment in my calculations. There is no such thing as "pure" anything, in racing, unless the horse is Secretariat of course. He's about as close to "pure" class as I've ever seen anyway.

I've said many times here that current form determines everything, class, speed, pace capabilities, distance capabilities, etc..

Class, IMO, cannot be used by itself, in a traditional class determination method, if your goal is to make long term profit. However, by including other factors in a performance-class method, in conjunction with value, can show long term profit.

I might mention that my "class" algorithm was created yesterday, in about 2 hours, strictly for this thread because of Al's invitation to post class ratings. I do not use this in my play. "Class", is included in my play through other calculated factors. In my program, there are "class" ratings, for purely comparison purposes, but they are strictly earnings based ratings, at qualified distances, and play no part in the picks produced by the program. They work well in some races and poorly in others, depending on other factors of the field and the race.

[Robert Fischer]

Are these class ratings or picks?

those are class ratings, and they are on the same scale as the bris speed figures.

[traynor]

I look forward eagerly to the explanations--pre-race or post-race--that address the thread topic. Namely, "How do you determine class?" NOT "Lookie me, lookie me, I got a secret and if you give me money maybe I'll tell you more."

I don't think anyone really cares whether the "posted selections" win or lose, because I don't think all that many people are willing to bet on other people's choices. If they are willing to do so, there are much better sources than freebie postings. I think the major interest is not which horse was selected in some specific race, but HOW it was selected, and WHY.

The thread was (and is) intended to trade information and ideas on the topic of methods for defining, measuring, and/or evaluating class.

[Capper Al]

There you go! :ThmbUp:

No tote-board was a condition. I was thinking why you might not see the importance of this. Programmers are like alchemist hunting for those magically numbers that will change their formulas into gold. The action is different from Paper and Pencil cappers.

[DeltaLover]

I don't think anyone really cares whether the "posted selections" win or lose, because I don't think all that many people are willing to bet on other people's choices.

I think the major interest is not which horse was selected in some specific race, but HOW it was selected, and WHY.



Exactly.

I have to add that any selection than is been made before the pools are open suffers from input shortage and should not be taken seriously; I never pay any attention to what public handicappers are suggesting, either in the TV or news papers and frankly I find it very strange that there people who still are using them as their main decision making engine!

The WHY a horse is selected most of the times requires an answer that involves the crowd's behavior and is tangible to the HOW (of course!)

Except special cases (like TC or BC events) I try not to study any race until the pools are open. I usually just eyeball tomorrow's program but nothing more than this. I have found this, to be my best approach for my style of handicapping and betting, that can be summarized in a handful of principles, very simple pace calculations and more than anything else application of game theory, something that has become the cornerstone of my approach.

[Capper Al]

I look forward eagerly to the explanations--pre-race or post-race--that address the thread topic. Namely, "How do you determine class?" NOT "Lookie me, lookie me, I got a secret and if you give me money maybe I'll tell you more."

I don't think anyone really cares whether the "posted selections" win or lose, because I don't think all that many people are willing to bet on other people's choices. If they are willing to do so, there are much better sources than freebie postings. I think the major interest is not which horse was selected in some specific race, but HOW it was selected, and WHY.

The thread was (and is) intended to trade information and ideas on the topic of methods for defining, measuring, and/or evaluating class.

Follow the discussion. We talked about what we did. Then, putting some action behind our words, we posted. So now that we laid our cards down, you are insulting us? We can add insulting to your other characteristics, along with arrogant and a know it all. I and others honestly participated in good faith in this thread.

[Robert Fischer]

So now that we laid our cards down, you are insulting us?

I didn't get insulting from what Traynor typed, but I'll leave it up to him to clarify.

Emphasis on the process could be interesting.

Everyone who posted class numbers did so in good faith, but even so it shouldn't be expected for us to reveal our methodology.

Whatever we do feel comfortable discussing, specific, or general is a bonus.

Just part of the give and take of discussing our approach to the game IMHO.

[Capper Al]

I didn't get insulting from what Traynor typed, but I'll leave it up to him to clarify.

Emphasis on the process could be interesting.

Everyone who posted class numbers did so in good faith, but even so it shouldn't be expected for us to reveal our methodology.

Whatever we do feel comfortable discussing, specific, or general is a bonus.

Just part of the give and take of discussing our approach to the game IMHO.

It's almost like one needs to ignore Traynor in his own thread. I'm glad you agree that the posts were in good faith. We'll see what happens tomorrow.

[raybo]

There's a chance for rain Saturday in Chicago, so here's today's card for Belmont. Same understanding that Al stated, 1 card means very little in the long term, so, no guarantees, play these at your own risk! I will post the top 2 class rated horses (my own auto-calculated performance class ratings), unless there are virtual ties, in which case I will post more than 2.

I will say that these ratings include overall performance and class in competitive races, and an adjustment for form cycle and current form. There are no subjective judgements made by me, strictly computer generated.

BEL 06/20/2013

Race 1 8.50f InnerT SOC
1 142.8 ---------3rd @ 6.5/1 odds
8 138.0
7 137.7

Race 2 6.00f Dirt Clm 30000
2 126.8 -----------2nd @ 5.8/1 odds
8 126.0

Race 3 6.00f Dirt Clm 14000
5 141.3 -------------1st @ 0.90/1 odds --- $3.80
7 135.3

Race 4 7.00f Turf SOC
3 281.4
5 153.9 -----------2nd @ 4.60/1 odds

Race 5 8.00f Dirt OClm 14000nw1
4 140.0
1 137.7

Race 6 8.50f InnerT Clm 20000
12 125.5
7 122.4 -----------4th @ 4.90/1 odds

Race 7 8.00f Turf SOC
8 140.9 ---------- 1st @ 4.60/1 odds ---$11.20
6 140.3 ---------2nd @ 2.85/1 odds

Race 8 8.00f Dirt OClm 25000nw2
7 157.8 -----------1st @ 2.15/1 odds ----$6.30
10 153.7 ----------- 3rd @ 5,40/1 odds
4 153.3

Race 9 7.00f Turf Md Sp Wt
3 133.0
5 131.0 -----------1st @ 4.40/1 odds ---- $ 10.80
2 130.0 -----------2nd @ 4.00/1 odds

4 winners --- 44.4% hit rate
4 places
2 shows
1 4th

$42 bet
$32.10 paid
$9.90 lost

I stated what the algorithm consists of earlier. Not a good day but not a complete bomb either. Take it for what it is, an incomplete handicapping approach which will work well some days and not so well other days. I'm certainly not embarrassed by the results. I'm pretty confident that if this class method (maybe tweaked a bit for race type, surface, track, etc..) were included with some other factors, and some value requirements, one could do ok long term.

[Capper Al]

Good job Raybo. 4 winners out of 9 without handicapping. Sometimes it seems that handicapping gets me off more winners than on.

[raybo]

Good job Raybo. 4 winners out of 9 without handicapping. Sometimes it seems that handicapping gets me off more winners than on.

It can do that, for sure. I think the key with handicapping, and betting, is consistency in approach in your handicapping, and discipline and patience in passing bad races and requiring value for the investment. It's the same old thing regardless of one's method.

[raybo]

Follow the discussion. We talked about what we did. Then, putting some action behind our words, we posted. So now that we laid our cards down, you are insulting us? We can add insulting to your other characteristics, along with arrogant and a know it all. I and others honestly participated in good faith in this thread.

Yeah, he's so afraid that someone's going to show something that might disprove some of his own theories, or that the sole purpose for someone posting is about promoting some product or offering to the members here (for money or for free), and at the same time, he posts lots of stuff that sounds an awful lot like self promotion, immature put downs, and enthusiastic ego stroking, it's just really sad, and angering.

[MJC922]

Speaking for my own ratings, from the perspective of a subscriber they're a black box. Some people will have a problem with that and they should. However sharing any proprietary info with respect to the calculations would not be a good idea for a lot of reasons and I don't think anyone here expects that from anyone else.

More important to many I believe, and this goes for any ratings one might use or even purchase, is to be in agreement conceptually with both the usefulness of what it is we're trying to quantify, including the various inputs. For the ratings I generate, the inputs are who beat who, by how much, and at what weight. Surfaces types (dirt turf and synthetics) play into things as well but aside from that it's a very basic premise when looked at from a high level.

Now the low level stuff on the other hand, the years of research, the implementation, the software, hardware, and the thousands of lines of code behind all of it, good luck to anyone getting anywhere with that. If they do manage to get anywhere near it they deserve whatever good comes out of it. Certainly those on this board toiling manually to get a line on class by going through charts closely deserve their rewards also. While I would like very much to post ratings before the races as others have I'm going to have to opt out for now, I believe the provider posts them the day after FWIW.

[Maximillion]

Interesting thread....but in my opinion,just posting each individuals "class ratings" would seem to lead to a dead end road,or a case of "my numbers are better than yours".Does it really help any of our fellow PA members?

I certainly dont expect anyone here to give away the candy store....but only speaking for myself......sometimes just a verbal opinion (well thought out) may be more accessible to many and create even more discussion.

[raybo]

Speaking for my own ratings, from the perspective of a subscriber they're a black box. Some people will have a problem with that and they should. However sharing any proprietary info with respect to the calculations would not be a good idea for a lot of reasons and I don't think anyone here expects that from anyone else.

More important to many I believe, and this goes for any ratings one might use or even purchase, is to be in agreement conceptually with both the usefulness of what it is we're trying to quantify, including the various inputs. For the ratings I generate, the inputs are who beat who, by how much, and at what weight. Surfaces types (dirt turf and synthetics) play into things as well but aside from that it's a very basic premise when looked at from a high level.

Now the low level stuff on the other hand, the years of research, the implementation, the software, hardware, and the thousands of lines of code behind all of it, good luck to anyone getting anywhere with that. If they do manage to get anywhere near it they deserve whatever good comes out of it. Certainly those on this board toiling manually to get a line on class by going through charts closely deserve their rewards also. While I would like very much to post ratings before the races as others have I'm going to have to opt out for now, I believe the provider posts them the day after FWIW.

I agree, as you said, from higher level - looking down, the factors are basic stuff, but that's only because all the time and research into them has already been accomplished, so you know what those factors are telling you. It's then pretty simple to put some of them together, using a little common sense, to accomplish a specific purpose, like a class rating for instance.

[MJC922]

4 winners --- 44.4% hit rate
4 places
2 shows
1 4th

$42 bet
$32.10 paid
$9.90 lost

I stated what the algorithm consists of earlier. Not a good day but not a complete bomb either. Take it for what it is, an incomplete handicapping approach which will work well some days and not so well other days. I'm certainly not embarrassed by the results. I'm pretty confident that if this class method (maybe tweaked a bit for race type, surface, track, etc..) were included with some other factors, and some value requirements, one could do ok long term.

Nice job. I looked through my ratings just grabbed the best fig last race to be quick about it. Last two races on the card came through on top. Last race was nice, Palace Dreams, one lifetime start, top fig, pays $10.80. Sometimes there's only one way to play a race and that last one would be all me. Unfortunately I was stuck in my cubicle looking forward to the weekend.

[raybo]

Interesting thread....but in my opinion,just posting each individuals "class ratings" would seem to lead to a dead end road,or a case of "my numbers are better than yours".Does it really help any of our fellow PA members?

I certainly dont expect anyone here to give away the candy store....but only speaking for myself......sometimes just a verbal opinion (well thought out) may be more accessible to many and create even more discussion.

Of course you're right to a certain extent about the postings, but just talking about something, conceptually, without providing examples, does not lend much credence to your conceptual discussions. That's why I stated the factors included in the ones I posted. The "recipe" for their creation are not necessary, that's for those interested in the approach to determine themselves. My belief is that if you don't at least do some testing and research yourself, you will never develop confidence in your ratings. And, in this game, if you don't have confidence in the viability of what you're betting on, you better not bet because you will not be consistent in your method or your betting and you'll be constantly second guessing yourself, and you will lose money, bottom line.

[raybo]

Nice job. I looked through my ratings just grabbed the best fig last race to be quick about it. Last two races on the card came through on top. Last race was nice, Palace Dreams, one lifetime start, top fig, pays $10.80. Sometimes there's only one way to play a race and that last one would be all me. Unfortunately I was stuck in my cubicle looking forward to the weekend.

I never even looked at the horse's PPs for that card, so I didn't know how many starts the horse had. I just posted the ratings as they were displayed in my race summary report. As I said, if you know what the ratings are telling you, you don't need to know all the exact percentages of their ingredients/factors because you created them in the first place. That's the importance of doing your own work and research, IMO. You can buy a house or you can build one yourself, which way gives you more information about the house?

[traynor]

And that is the explanation of how one determines class in a thoroughbred race horse.

[traynor]

I had hoped for a bit more openness, and a bit less obfuscation. Bluntly, I don't think anyone has anything worth hiding, and attempts to be coy about specifics adds little. Lots of people have spent lots of time trying various combinations of things in attempts to quantify class. Round things that roll are not a miracle. The wheel was invented long ago, and it is not necessary to endlessly re-hash the same approaches that thousands of others have tried, used, and discarded as hopelessly inadequate.

No "formnulas" are expected (or needed, or wanted, because basically I think the formulas being used--at least by those posting on the Internet, on this forum and others--are little more than variations on the same dull, boring concepts that have been hammered into oblivion by thousands of others computer and pen-and-paper handicappers over decades).

As I said, it is irrelevant whether the "selections" won or lost. I'm curious what process was/is used in your particular approach to determining class. That's all. I don't want your "secret formulas," for the reasons stated above.

[raybo]

I had hoped for a bit more openness, and a bit less obfuscation. Bluntly, I don't think anyone has anything worth hiding, and attempts to be coy about specifics adds little. Lots of people have spent lots of time trying various combinations of things in attempts to quantify class. Round things that roll are not a miracle. The wheel was invented long ago, and it is not necessary to endlessly re-hash the same approaches that thousands of others have tried, used, and discarded as hopelessly inadequate.

No "formnulas" are expected (or needed, or wanted, because basically I think the formulas being used--at least by those posting on the Internet, on this forum and others--are little more than variations on the same dull, boring concepts that have been hammered into oblivion by thousands of others computer and pen-and-paper handicappers over decades).

As I said, it is irrelevant whether the "selections" won or lost. I'm curious what process was/is used in your particular approach to determining class. That's all. I don't want your "secret formulas," for the reasons stated above.

I already gave my thinking regarding the factors I used. If that isn't enough then I suggest you figure it out yourself. Oh wait, you obviously already know how to do it, so why are you still asking, just to start a conversation? Mine aren't for sale by the way, not because they should be protected but because it is an incomplete method, class being only part of the equation, and not worth anything by itself, even for free.

By the way, you are the one who said "50% winners with 2 picks" would be significant for a single card, not that it's "irrelevant whether the "selections" won or lost". So, which is it?

[rubicon55]

IMHO, maybe I am rehashing what has been said since the beginning of time but I ask myself the question who has the horse in question beat and who has beaten him/her and under what conditions and for what kind of purse value. I discount dream trips and heavy biases and look for what I think are honest efforts and give bonuses for those who have overcome the pace, traffic and had to race wide. I imagine there could be a good discussion if lengths mattered at the finish in the final analysis. I cannot seem to find a method that works best for me that would assign on paper or a computer a number or a value that would be called class value. A single number to me is too static for me in a race dynamic. Too many variables confuses regarding the definition of class is not useful to me. I find it more important in observing how a horse rates during a race IMO which can be a better indicator of ability at different purse levels, this may hold the keys if that horse can run competitively at that purse level or not. We all can tell by the first or second call who rating and who does not. I pay more attention where my runner is at the first call and how the race unfolds between the first and second call (sound familiar?) If the horses rates and does not run rank or whither under pressure then I consider that is the class level of that horse for this race until today's race is run in which he may rise or fall per my definition of class rating. Seen too many trainers putting horses at the wrong purse level. Obviously the horse in question need to finish well, say at least mid pack at the finish unless it is a blanket finish just to make sure he is not a quitter type. It is sort of a hybrid of trip handicapping supported by pace, speed and past performance (form cycle) under similar purse conditions. I start with my contenders and then start reviewing. Takes a lot of time to review videos and not convenient like a computer but not sure how else I can separate the chaff from the wheat. As other posters have stated the charts are only as good as the one making them. This method of course limits the races I can cap due to the review time so I have to be selective - in that case like most folks I go to the tracks, surfaces, distances and purses that I do better at. Very difficult to define class IMO.

[traynor]

IMHO, maybe I am rehashing what has been said since the beginning of time but I ask myself the question who has the horse in question beat and who has beaten him/her and under what conditions and for what kind of purse value. I discount dream trips and heavy biases and look for what I think are honest efforts and give bonuses for those who have overcome the pace, traffic and had to race wide. I imagine there could be a good discussion if lengths mattered at the finish in the final analysis. I cannot seem to find a method that works best for me that would assign on paper or a computer a number or a value that would be called class value. A single number to me is too static for me in a race dynamic. Too many variables confuses regarding the definition of class is not useful to me. I find it more important in observing how a horse rates during a race IMO which can be a better indicator of ability at different purse levels, this may hold the keys if that horse can run competitively at that purse level or not. We all can tell by the first or second call who rating and who does not. I pay more attention where my runner is at the first call and how the race unfolds between the first and second call (sound familiar?) If the horses rates and does not run rank or whither under pressure then I consider that is the class level of that horse for this race until today's race is run in which he may rise or fall per my definition of class rating. Seen too many trainers putting horses at the wrong purse level. Obviously the horse in question need to finish well, say at least mid pack at the finish unless it is a blanket finish just to make sure he is not a quitter type. It is sort of a hybrid of trip handicapping supported by pace, speed and past performance (form cycle) under similar purse conditions. I start with my contenders and then start reviewing. Takes a lot of time to review videos and not convenient like a computer but not sure how else I can separate the chaff from the wheat. As other posters have stated the charts are only as good as the one making them. This method of course limits the races I can cap due to the review time so I have to be selective - in that case like most folks I go to the tracks, surfaces, distances and purses that I do better at. Very difficult to define class IMO.

What you are doing is a lot like what I am doing. I think one of the important factors--as you point out above--is that there are often multiple"real" class levels within a conventional class level (often at identical or similar purses) that can be separated by race conditions. and by the composition of the horses competing in the individual races.

In essence, it is more useful to compare a horse against the other horses it was racing against, rather than using artificial categories like purse or claiming price. At some tracks there are multiple grades within a given class that create a hierarchy of horses that cycle through according to their current fitness and condition. Viewed superficially, or with simplistic categorizations, it seems the horse are all "running at the same class level." That is a long way from being the case in the real world.

[Capper Al]

Next couple of days Arlington is having thunderstorms. I avoid off tracks if I can. All my stats are based on fast and firm tracks.

As far as formulas for class go, there are 3 basic methods that come to mind:


Earnings per start (ESP)
Average purchase value (APV)
Who beat who on a per race basis
Many books talk about these and provide stats for the first two. I am sure if most of us opened our code we would find that we are doing similar things.

[illinoisbred]

Al..the poly is always labeled fast. It does not become "off" when wet. If anything,if substantial rain hits,it will play faster. Seeya Tomorrow!

[Capper Al]

Al..the poly is always labeled fast. It does not become "off" when wet. If anything,if substantial rain hits,it will play faster. Seeya Tomorrow!

It's better than dirt. The only big thing to watch for then, when it comes to rain, is the turf to poly. Looking forward to seeing you.

[pandy]

The toughest thing is evaluating how good the fields were that the horse raced in. It requires not only keeping tabs on key races but knowing what types of races the winners from those key races won. If you have a computer system that can rate class in this manner you should get a good class rating.

Speed ratings are often used as part of a formula to create class ratings but are not good for this purpose for many reasons.

[traynor]

Next couple of days Arlington is having thunderstorms. I avoid off tracks if I can. All my stats are based on fast and firm tracks.

As far as formulas for class go, there are 3 basic methods that come to mind:

Earnings per start (ESP)
Average purchase value (APV)
Who beat who on a per race basis
Many books talk about these and provide stats for the first two. I am sure if most of us opened our code we would find that we are doing similar things.

One of the most useful class ratings using most of the factors you mention above was a compound: APV is divided by today's purse to get a percentile relationship--like 0.95 or 1.12 or whatever. That value is then combined with win% and/or win-place percentage to establish the basic figure. The formula I learned years ago included EPS, but in "layers"--meaning the resulting "rating" could not directly compare $2500 claimers to allowance runners, but was extremely accurate within either (or any other) type of race. To adjust EPS by class level takes a lot of record keeping, to establish an average for EPS at that class level at that track. That average value is set to "1.0" and the actual EPS are divided by the average EPS to get another value similar to APV figure (0.95 or 1.12 or whatever).

The EPS values are important--unless they are grade/track specific, they will tweak the resulting rating all over the place.

[traynor]

The toughest thing is evaluating how good the fields were that the horse raced in. It requires not only keeping tabs on key races but knowing what types of races the winners from those key races won. If you have a computer system that can rate class in this manner you should get a good class rating.

Speed ratings are often used as part of a formula to create class ratings but are not good for this purpose for many reasons.

I might add that it is also important to evaluate the races other than key races, for the same purpose. That is, to establish some means of categorizing each race based on the relative strength or weakness of the entries in that race. There is a lot of work involved (at least initially) but the results can be impressive.

Most bettors look at a set group of values for comparing performances. In general, a $7500 claimer is considered a $7500 claimer, and that is the end of it. Exactly as in the key race example you mention, the field in one $7500 claimer may be tough as nails, and in another--same class level, same purse, same track--be a cake walk. Knowing which are key races is important. Just as important is knowing which are "soft" or "average" races, for subsequent evaluations of the entries that competed against each other in those races.

Again, it takes a LOT of work, but the efforts can be well rewarded.

[Capper Al]

One of the most useful class ratings using most of the factors you mention above was a compound: APV is divided by today's purse to get a percentile relationship--like 0.95 or 1.12 or whatever. That value is then combined with win% and/or win-place percentage to establsihd the basic figure. The formula I learned years ago included EPS, but in "layers"--meaning the resulting "rating" could not directly compare $2500 claimers to allowance runners, but was extremely accurate within either (or any other) type of race. To adjust EPS by class level takes a lot of record keeping, to establish an average for EPS at that class level at that track. That average value is set to "1.0" and the actual EPS are divided by the average EPS to get another value similar to APV figure (0.95 or 1.12 or whatever).

The EPS values are important--unless they are grade/track specific, they will tweak the resulting rating all over the place.

So you know a little something about class.

[Capper Al]

The toughest thing is evaluating how good the fields were that the horse raced in. It requires not only keeping tabs on key races but knowing what types of races the winners from those key races won. If you have a computer system that can rate class in this manner you should get a good class rating.

Speed ratings are often used as part of a formula to create class ratings but are not good for this purpose for many reasons.

First and foremost about making class figs is not to use speed in anyway in the calculations. What is being attempted by a class fig is a little independance from speed figures. Even with no speed figures involved, they'll still correlate to each other.

Focusing on winners and key races gets into who beat who. This is hard to follow and make sense of.

[traynor]

So you know a little something about class.

The formula above was presented by Michael Pizzola at a Sartin seminar in New Orleans in the 1980s, in conjunction with a money management system used by a group of professional bettors in New York. According to Pizzola, they had developed their class rating to a point that they virtually ignored pace, speed, and all the rest--comparisons were made on the basis of their class ratings, with wagers made on the basis of "gaps" ("differences") between the ratings.

A group of professional bettors from Argentina hired me (way back) to code an app that provided a structured approach to including trip notes and body language observations in a "conventional" (speed/pace) application. They used a combination of factors to establish individual race ratings, that were in turn used somewhat like Beyer's projected speed figures to create an individual "race rating" that could be used to compare past performances from those races in subsequent races.

Both approaches were rewarding--at least at the time and for some years after. Much more sophisticated approaches are used now by serious bettors.

[rubicon55]

What you are doing is a lot like what I am doing. I think one of the important factors--as you point out above--is that there are often multiple"real" class levels within a conventional class level (often at identical or similar purses) that can be separated by race conditions. and by the composition of the horses competing in the individual races.

In essence, it is more useful to compare a horse against the other horses it was racing against, rather than using artificial categories like purse or claiming price. At some tracks there are multiple grades within a given class that create a hierarchy of horses that cycle through according to their current fitness and condition. Viewed superficially, or with simplistic categorizations, it seems the horse are all "running at the same class level." That is a long way from being the case in the real world.

On this note, I saw somewhere a chart that "normalized" purses between tracks. The A Class tracks I think was the baseline and the mid class and minor tracks were normalized to balance out the actual purse values so each track could be compared. Do you recall seeing such a table? I am not sure if it was Cynthia Publishing, HANA or what. The tracks were given a numerical value and not based on purse value. This would be especially handy comparing invaders with questionable purse earnings. Can anybody comment?

[traynor]

First and foremost about making class figs is not to use speed in anyway in the calculations. What is being attempted by a class fig is a little independance from speed figures. Even with no speed figures involved, they'll still correlate to each other.

Focusing on winners and key races gets into who beat who. This is hard to follow and make sense of.

I could not agree more. The advantage is that discrete (separate) figures can then be "meaningfully combined" in ways that are impossible with "mixed" values. That incorporates the notion of "confounding variables" that is thoroughly muddled with combined (speed and class, pace and class) figures. Specifically, how predictive is A when B is => than X, and how predictive is A when B is <= Y--and, most interesting of all, how predictive is A when B is not factored in at all?

I think it is possible to derive much more predictive (and much more profitable) values by keeping the factors of pace, speed, class, current form, etc. completely separate.

[traynor]

On this note, I saw somewhere a chart that "normalized" purses between tracks. The A Class tracks I think was the baseline and the mid class and minor tracks were normalized to balance out the actual purse values so each track could be compared. Do you recall seeing such a table? I am not sure if it was Cynthia Publishing, HANA or what. The tracks were given a numerical value and not based on purse value. This would be especially handy comparing invaders with questionable purse earnings. Can anybody comment?

I have seen a number of such charts, (including the DRF), but all seemed based on comparisons of "average purse values" by track, with the implication that track-to-track class "adjustments" can be made using such a chart. I stopped trying to use them years ago, and decided it would be better (for me) to develop my own approach to individual race class rather than track class. That seemed (seems) to work better than track comparisons.

For example, on the southwest circuit, bottom level claimers at one track are "better" than allowance competitors at another track, simply because of the competition in the respective races. Purses and earnings tend to present a very different (and much less accurate) view of that competition.

[traynor]

First and foremost about making class figs is not to use speed in anyway in the calculations. What is being attempted by a class fig is a little independance from speed figures. Even with no speed figures involved, they'll still correlate to each other.

Focusing on winners and key races gets into who beat who. This is hard to follow and make sense of.

Not as hard as you might think. It presents a problem if you apply it to individual races going forward. It is much less difficult if you handicap races every day (whether you bet on them or not) and generate class ratings for all the entries, that can in turn be used to create a "race class" rating for each of the races.

It takes some work, but the values generated are worthwhile. I sometimes forget that people still handicap on the basis of "fresh" data. The app(s) I use create files of "handicapped races" that are used to handicap subsequent races. For example, I ignore any race with less than three contenders (according to very specific criteria) as "noise" that does little more than introduce nonsense into otherwise useful data.

The reason--a race with one or two standouts is overbet on the front end. It is also unlikely that a "lesser" entry in that race--that would otherwise have a good shot at winning or placing--will do more than use the race as an extended workout, saving its run for a softer spot. There are dozens of similar scenarios that make uncritical acceptance of "all past races as equally significant" very hard on the bottom line. Class ratings--and, by extension--class rating of races based on those class ratings--could go a long way to improving that situation.

[traynor]

If all that "race class" stuff seems to hard to handle, try an experiment. If you have a class rating figure, assign that figure to (calculate it for) every entry in a race. Then divide the sum by the number of entries. Call that the "average class of the race." That approach is a little simplistic, but if you tinker around a bit, you can come up with values that accurately rate the "class" of each race--that have absolutely nothing to do with speed or pace.

If the class ratings are accurate going in, the values coming out will be much more useful for comparisons than the artificial purse levels. Obviously, an entry that ran well against a "3.76" (or whatever) field did more than an entry that ran well against a "2.89" (or whatever) field. The operant condition is that the class ratings are accurate going in, and actually represent the factor of "class."

[DeltaLover]

I can see that a good percentage of you are insisting in the concept of a 'class rating' expressed in numerical scale as a speed figure.

Although I do not believe such a metric can add any value to the bettor, I have to underline that such a rating is easy to to constructed (and most likely this is exactly why it adds no value).

The most important concept when creating a metric, is how to evaluate it. Once some handicapping attribute has been converted to a numeric quantity, its 'speciality' automatically becomes irrelevant. With this I mean that regardless of 'what' your figure is measuring it should be regarded as just a 'clear' (dimensionless) metric and nothing more. Such a metric can easily be compared to another and although in this posting I will not delve in to details conclusions can be drawn to decide which one is more effective. Following such an approach we can prove that the 'best' metrics are exactly those who replicate the behavior of the crowd; in other words if we were trying to create the ultimate figure having unlimited resources both in data and in computing power, chances are that we would end up with a composite encapsulating the fundamentals resulting to very high correlation with the crowd's odds. This is why effort consumed towards this direction has more academic importance than real betting value. As an example of this I refer you to this document:

http://eprints.usq.edu.au/4300/1/Williams_Li_Publ_version.pdf


Many years ago, when starting to apply pattern recognition and AI to race prediction I had discovered this document among other similar and replicated and improved their performance, only to find out that what I was doing was of no real value, as it would have been easier to reach the same conclusions just by eyeballing the board...

Fortunately, horse betting is not that simple. I say fortunately because otherwise everyone would follow an optimal approach quickly evaporating any possibility for profits!

If you really insist in class rating (or whatever else kind of metric) what you should concentrate is not the algorithm to use like for example:



Earnings per start (ESP)
Average purchase value (APV)
Who beat who on a per race basis




as all these factors represent 'handicapping 101' material that will never give you any betting edge.

What might give you the edge, is to try to estimate what the probability and margin of error in the calculation of the figure.

This is the only estimate giving you some edge that you can transform to profits.

Needless to say that exactly the same applies not only to class but any other numerical figure you might be using....

[traynor]

I can see that a good percentage of you are insisting in the concept of a 'class rating' expressed in numerical scale as a speed figure.

Although I do not believe such a metric can add any value to the bettor, I have to underline that such a rating is easy to to constructed (and most likely this is exactly why it adds no value).

The most important concept when creating a metric, is how to evaluate it. Once some handicapping attribute has been converted to a numeric quantity, its 'speciality' automatically becomes irrelevant. With this I mean that regardless of 'what' your figure is measuring it should be regarded as just a 'clear' (dimensionless) metric and nothing more. Such a metric can easily be compared to another and although in this posting I will not delve in to details conclusions can be drawn to decide which one is more effective. Following such an approach we can prove that the 'best' metrics are exactly those who replicate the behavior of the crowd; in other words if we were trying to create the ultimate figure having unlimited resources both in data and in computing power, chances are that we would end up with a composite encapsulating the fundamentals resulting to very high correlation with the crowd's odds. This is why effort consumed towards this direction has more academic importance than real betting value. As an example of this I refer you to this document:

http://eprints.usq.edu.au/4300/1/Williams_Li_Publ_version.pdf


Many years ago, when starting to apply pattern recognition and AI to race prediction I had discovered this document among other similar and replicated and improved their performance, only to find out that what I was doing was of no real value, as it would have been easier to reach the same conclusions just by eyeballing the board...

Fortunately, horse betting is not that simple. I say fortunately because otherwise everyone would follow an optimal approach quickly evaporating any possibility for profits!

If you really insist in class rating (or whatever else kind of metric) what you should concentrate is not the algorithm to use like for example:



as all these factors represent 'handicapping 101' material that will never give you any betting edge.

What might give you the edge, is to try to estimate what the probability and margin of error in the calculation of the figure.

This is the only estimate giving you some edge that you can transform to profits.

Needless to say that exactly the same applies not only to class but any other numerical figure you might be using....

The key point is that the paper mentioned is based on a "study" of 143 races in Jamaica in 2007. With a small enough (and selectively selected enough) sample, pretty much anything can be proven statistically to be "true."

I agree that recycling the same old same old is pretty much useless as far as profit is concerned. I think the first step in a better direction is the creation of individual (unique, proprietary) "ratings" that are NOT simply variations on the same old, worn out, overused approaches that everyone else is using. I think a major stumbling block in that understanding is the realization that one has to start somewhere--and readily available data is that starting place. How it is used, how it is interpreted, how it is separated and combined in new and creative ways, is what makes the result useful.

The result may well be the conclusion that doing the same old same old is worthless. It was not that many years ago that the sharpies were declaring, "Look, Matilda, the three horse is carrying three pounds less this race. Quick, bet the rent money to win on the the three!" Or some equally absurd nonsense involving the "average of the last two Beyer's," or APV, or whatever. If one assumes that infinite wisdom is contained in the downloaded data, one tends to chase rainbows looking for insights that don't exist. Understanding that deficiency enables one to look elsewhere, or to look at the same data in different ways. Again, one has to start the process somewhere, and the ordinary (and mostly useless) conventional approaches to defining "class" are as good a place as any.

I do not think it is especially useful to declare all metrics as useless. That may indicate nothing more than that one has not discovered the appropriate metric(s). It does not in any way establish that such metrics do not exist (they obviously do, and are used successfully on a daily basis), or that attempts to create, discover, or calculate such metrics are misguided.

[raybo]

The key point is that the paper mentioned is based on a "study" of 143 races in Jamaica in 2007. With a small enough (and selectively selected enough) sample, pretty much anything can be proven statistically to be "true."

I agree that recycling the same old same old is pretty much useless as far as profit is concerned. I think the first step in a better direction is the creation of individual (unique, proprietary) "ratings" that are NOT simply variations on the same old, worn out, overused approaches that everyone else is using. I think a major stumbling block in that understanding is the realization that one has to start somewhere--and readily available data is that starting place. How it is used, how it is interpreted, how it is separated and combined in new and creative ways, is what makes the result useful.

The result may well be the conclusion that doing the same old same old is worthless. It was not that many years ago that the sharpies were declaring, "Look, Matilda, the three horse is carrying three pounds less this race. Quick, bet the rent money to win on the the three!" Or some equally absurd nonsense involving the "average of the last two Beyer's," or APV, or whatever. If one assumes that infinite wisdom is contained in the downloaded data, one tends to chase rainbows looking for insights that don't exist. Understanding that deficiency enables one to look elsewhere, or to look at the same data in different ways. Again, one has to start the process somewhere, and the ordinary (and mostly useless) conventional approaches to defining "class" are as good a place as any.

I do not think it is especially useful to declare all metrics as useless. That may indicate nothing more than that one has not discovered the appropriate metric(s). It does not in any way establish that such metrics do not exist (they obviously do, and are used successfully on a daily basis), or that attempts to create, discover, or calculate such metrics are misguided.

Yup, yup, and yup. :ThmbUp:

[DeltaLover]

The key point is that the paper mentioned is based on a "study" of 143 races in Jamaica in 2007. With a small enough (and selectively selected enough) sample, pretty much anything can be proven statistically to be "true.

Sure. I can also add that even a relatively large sample can spawn erroneous conclusions, especially when outliers are not handled properly. Besides that, what I wanted to underline with this example, was my opinion that such an approach, might present more academic rather than betting interest.

I do not think it is especially useful to declare all metrics as useless

Not necessary, but I think it is of prime importance to develop a method to allow to make such a declaration.

This 'method' can be seen as a 'meta' metric, a metric or metrics, implementing a tool allowing the user to declare a metric as useless or not!


That may indicate nothing more than that one has not discovered the appropriate metric(s). It does not in any way establish that such metrics do not exist (they obviously do, and are used successfully on a daily basis)

I think that the appropriate metric(s) belong in the meta - metric ring rather than first degree metric(s) (like speed figures for example). What is of imperative importance for the bettor is the behavior of the fundamental metrics and this exactly what we really need to measure.


, or that attempts to create, discover, or calculate such metrics are misguided.


Having a functional set of meta-metrics, you no longer need to worry about misguided metrics since they will be recognized and either improved or rejected as pure random filters..

[traynor]

Sure. I can also add that even a relatively large sample can spawn erroneous conclusions, especially when outliers are not handled properly. Besides that, what I wanted to underline with this example, was my opinion that such an approach, might present more academic rather than betting interest.



Not necessary, but I think it is of prime importance to develop a method to allow to make such a declaration.

This 'method' can be seen as a 'meta' metric, a metric or metrics, implementing a tool allowing the user to declare a metric as useless or not!



I think that the appropriate metric(s) belong in the meta - metric ring rather than first degree metric(s) (like speed figures for example). What is of imperative importance for the bettor is the behavior of the fundamental metrics and this exactly what we really need to measure.



Having a functional set of meta-metrics, you no longer need to worry about misguided metrics since they will be recognized and either improved or rejected as pure random filters..

I mostly agree with your observations. One thing that I believe creates a LOT of confusion in research studies (not just in regard to horse racing) is the influence of confounding variables. In an oversimplification to demonstrate the principle to the casual reader, speed and pace figures (for example) that ignore the effects of class, jockey, trainer, weather, or any one of dozens (or hundreds or thousands) of other factors that may exert varying degrees of influence on the values studied will routinely come up short in measuring what they are assumed to be measuring. Values/ratings/figures derived from such research will be conceptually flawed out of the gate, because important factors were "overlooked" or "ignored." Discovering and studying those "important factors" is something few want to do--so they build larger databases in the hope that it will all somehow "even out." Nice in theory, but not so good when applied in the real world to situations one is wagering on.

I agree completely that what you refer to as first degree metrics are mostly useless (for profit). Doing the same old same old that everyone else is doing, using data from the same information silo(s), is little more than busywork that accomplishes little of value.

[raybo]

Continuing with the class figures we were posting yesterday. Al had suggested AP for today and Saturday, the track may be wet and off the turf but we'll see what happens anyway. I'm also continuing with Belmont today. Top 2 class rated horses unless there is a huge standout (1 will be posted), or a virtual tie in which case more than 2 will be posted. The ratings include a purse value metric, from competitive races only, a performance metric, and a form cycle/current form adjustment. Really pretty basic stuff, even the performance rating. Form cycles are 1-8 but only level 1 and 2 are considered for adjustments to the rating, also improvement/decline in form in the last 3 races add or subtract to/from the ratings (basically a "+", " -", or "=" notation is converted to +2, -2, or 0 points respectively). So, you can see this rating is not at all complex, compared to what some primarily class handicappers do. Don't know how an "off" track will affect these ratings so I will not address that scenario at all.

Also, no "'traditional" database research (unless you include past experience as a traditional database) was done regarding the creation of these ratings, just common sense. So, I guess the question is, how good is common sense in the creation of ratings, regardless of the type of ratings being created. Personally, I think common sense is of great importance in both handicapping methods and in wagering. Of course, any method that has been well designed is worthless if you have little patience, and/or discipline, and/or consistency in remaining true to the method.

APX 06/21/2013
Race 1 7.00f Dirt Clm 7500
2 91.6
1 83.5
4 82.5
Race 2 9.00f Dirt Clm 7500
5 89.8
3 89.0
Race 3 -8.00f Turf OClm 80000clf
2 350.8
Race 4 6.00f Dirt OClm 40000nw2
1 160.0
4 158.3
Race 5 -5.00f Turf Clm 25000
5 126.3
2 112.4
Race 6 7.00f Dirt Md 15000
1 103.0
3 86.5
Race 7 -5.00f Turf Md Sp Wt
1 118.0
8 114.5
Race 8 8.50f Dirt Clm 5000
1 136.2
4 108.1
Race 9 -5.00f Turf Clm 25000
6 130.7
3 122.8
9 122.0


BEL 06/21/2013
Race 1 8.50f InnerT SOC
1 164.1
6 155.9
Race 2 5.00f Dirt Md Sp Wt
Pass - No data
Race 3 5.00f Dirt Md Sp Wt
Pass - No data
Race 4 8.50f Turf Md 35000
9 134.7
4 118.3
Race 5 8.00f Dirt OClm 25000nw1
6 162.0
1 155.0
Race 6 8.50f Turf Clm 20000
8 126.0
1 120.8
9 120.0
Race 7 8.00f Dirt OClm 14000nw1
1A 158.0
8 148.0
1 147.0
Race 8 8.50f InnerT OClm 25000nw2
2 166.9
9 156.7
Race 9 8.50f Turf Md 20000
9 112.1
8 110.3

[raybo]

Note: Belmont race 1 - scratch 1,1A,4,9,and 10, so the 6 and 3 are the top 2 rated horses in that race.

[thaskalos]

I mostly agree with your observations. One thing that I believe creates a LOT of confusion in research studies (not just in regard to horse racing) is the influence of confounding variables. In an oversimplification to demonstrate the principle to the casual reader, speed and pace figures (for example) that ignore the effects of class, jockey, trainer, weather, or any one of dozens (or hundreds or thousands) of other factors that may exert varying degrees of influence on the values studied will routinely come up short in measuring what they are assumed to be measuring. Values/ratings/figures derived from such research will be conceptually flawed out of the gate, because important factors were "overlooked" or "ignored." Discovering and studying those "important factors" is something few want to do--so they build larger databases in the hope that it will all somehow "even out." Nice in theory, but not so good when applied in the real world to situations one is wagering on.

I agree completely that what you refer to as first degree metrics are mostly useless (for profit). Doing the same old same old that everyone else is doing, using data from the same information silo(s), is little more than busywork that accomplishes little of value.

I think you've touched on something very important here...which may explain why many handicappers are giving a long look to "class" at this point in time.

It is widely believed that speed/pace handicapping was very profitable in years past -- when the vast majority of the handicappers were "class/consistency" minded -- but that these profits largely dried up once the importance of speed/pace was revealed to the handicappers at large.

Could it be that people now consider CLASS to be the "forgotten" factor in handicapping...just as speed was 40 years ago?

Do they think that class handicapping might be as profitable now as speed handicapping supposedly was in the pre-Beyer years?

[raybo]

I think you've touched on something very important here...which may explain why many handicappers are giving a long look to "class" at this point in time.

It is widely believed that speed/pace handicapping was very profitable in years past -- when the vast majority of the handicappers were "class/consistency" minded -- but that these profits largely dried up once the importance of speed/pace was revealed to the handicappers at large.

Could it be that people now consider CLASS to be the "forgotten" factor in handicapping...just as speed was 40 years ago?

Do they think that class handicapping might be as profitable now as speed handicapping supposedly was in the pre-Beyer years?

Maybe it's just another "retro" thing! This phenomenon might point to the importance of having a "comprehensive" approach, and adjusting the overall method as required over time. Racing isn't static, so why should our methods remain static?

[rubicon55]

Maybe it's just another "retro" thing! This phenomenon might point to the importance of having a "comprehensive" approach, and adjusting the overall method as required over time. Racing isn't static, so why should our methods remain static?

I agree with you, Thaskalos and Traynor per your comments - developing a "class" principle could potentially develop into something that the public has ignored as of late and thus becomes profitiable angle when used in conjunction with proven already methods (that works for you) provided it is used and tested properly - especially when all things "appear" equal on the surface and difficult to make distinctions (i.e. such as chaos type races). Interesting food for thought. Thanks.

[raybo]

For those who believe in "inherent" class, Race 4 at Belmont might interest you. The 4 controlled the early pace to the stretch call, where he was overtaken by what appeared to be a strong closer. However, the 4 took the lead back before the line winning by about a length. Some horses just have that extra something inside them.

By the way, the 9 horse scratched out of that race, leaving me with the 4 and 2 as the top 2 class horses.

[iceknight]

How do you determine class in CD R7 - this was a turf race and the second favorite (:6: Big lucky, made a good stretch move from his stalking position, then drifted wide and was easily passed by two closers.
The winner Ridge City was a horse dropping down from ALW levels to 30k claimers, but he has run before in 20/15k and 8kn3l claimers.

Ridge city won at 40-1 and is the first winner in last 2 yrs for owner (http://www.equibase.com/profiles/Results.cfm?type=People&searchType=O&eID=786502) /trainer (http://www.equibase.com/profiles/Results.cfm?type=People&searchType=T&eID=13992)

Congratulations to him. I don't know this person, but it is interesting to see the stats of a trainer/owner who does it only occasionally. When we compare this person to some of the other operations, we get a more realistic picture of racing occupations.

[raybo]

Belmont Race 8 - scratch 1,1A ,9,11 leaving the top 2 class rated horses as 2 and 10.

[traynor]

I think you've touched on something very important here...which may explain why many handicappers are giving a long look to "class" at this point in time.

It is widely believed that speed/pace handicapping was very profitable in years past -- when the vast majority of the handicappers were "class/consistency" minded -- but that these profits largely dried up once the importance of speed/pace was revealed to the handicappers at large.

Could it be that people now consider CLASS to be the "forgotten" factor in handicapping...just as speed was 40 years ago?

Do they think that class handicapping might be as profitable now as speed handicapping supposedly was in the pre-Beyer years?

To say class is one of a number of overlooked (or less emphasized in favor of more popular approaches) factors might be more descriptive.

There are approaches (a more appropriate term than specific strategies or tactics) that are as useful now as speed (and then pace) have been in the past. The advantage is in the fact that speed and pace still produce enough winners (with enough work and more than a little luck) to keep bettors chasing the same receding mutuel pools. Because bettors have become "comfortable" using the old approaches, they continue using them because they "kind of work." Some of the time. On a good day. When the moon is full. When those bettors contemplate change, a little voice whispers in their head, "Don't try to fix it--it ain't broke."

That little voice is lying, folks. It is broken--not because it doesn't occasionally generate a profit, and sometimes even an impressive profit--but because an equal amount of work applied in another area/other areas might be substantially more rewarding.

Consider how many times "alternative ideas" have been posted on this forum (besides the conventional approaches to speed and class) and those ideas have been immediately pooh-poohed by those claiming "generous ROIs every year, year after year." Or "seven-figure incomes from churning at the rebate shops", with no clue whatsoever how those doing so are (supposedly) generating a high enough rate of return WITHOUT the rebates to make the rebates "profitable" rather than just enticement for the big losers to keep losing big.

"Class"--as an approach to race analysis--is a topic that is complex enough, and difficult enough, to offer substantial advantage to those willing to do the work necessary.

[raybo]

To say class is one of a number of overlooked (or less emphasized in favor of more popular approaches) factors might be more descriptive.

There are approaches (a more appropriate term than specific strategies or tactics) that are as useful now as speed (and then pace) have been in the past. The advantage is in the fact that speed and pace still produce enough winners (with enough work and more than a little luck) to keep bettors chasing the same receding mutuel pools. Because bettors have become "comfortable" using the old approaches, they continue using them because they "kind of work." Some of the time. On a good day. When the moon is full. When those bettors contemplate change, a little voice whispers in their head, "Don't try to fix it--it ain't broke."

That little voice is lying, folks. It is broken--not because it doesn't occasionally generate a profit, and sometimes even an impressive profit--but because an equal amount of work applied in another area/other areas might be substantially more rewarding.

Consider how many times "alternative ideas" have been posted on this forum (besides the conventional approaches to speed and class) and those ideas have been immediately pooh-poohed by those claiming "generous ROIs every year, year after year." Or "seven-figure incomes from churning at the rebate shops", with no clue whatsoever how those doing so are (supposedly) generating a high enough rate of return WITHOUT the rebates to make the rebates "profitable" rather than just enticement for the big losers to keep losing big.

"Class"--as an approach to race analysis--is a topic that is complex enough, and difficult enough, to offer substantial advantage to those willing to do the work necessary.

So, are you saying pace and form are not complex enough for you? :bang:

[iceknight]

the issue I have with "class" is that it is hard to quantify and distribute that concept..Is their one uniform definition for it? Also, when a horse stepping up in class, it is clearly hard to determine if that horse belongs to this class or not.. until he wins or comes a close 2nd etc.. now take this by the large number of horses that are constantly going up or down and throw in some other factors like weather/track surface condition/bumps/steadying etc - which may cause a legitimate horse at a certain class level to still finish well below that level. How can class be determined and applied uniformly as a parameter?
"you will know it when you see it" seems to work more with class?

[iceknight]

Next couple of days Arlington is having thunderstorms. I avoid off tracks if I can. All my stats are based on fast and firm tracks.

As far as formulas for class go, there are 3 basic methods that come to mind:


Earnings per start (ESP)
Average purchase value (APV)
Who beat who on a per race basis
Many books talk about these and provide stats for the first two. I am sure if most of us opened our code we would find that we are doing similar things. Thanks!

[rubicon55]

the issue I have with "class" is that it is hard to quantify and distribute that concept..Is their one uniform definition for it? Also, when a horse stepping up in class, it is clearly hard to determine if that horse belongs to this class or not.. until he wins or comes a close 2nd etc.. now take this by the large number of horses that are constantly going up or down and throw in some other factors like weather/track surface condition/bumps/steadying etc - which may cause a legitimate horse at a certain class level to still finish well below that level. How can class be determined and applied uniformly as a parameter?
"you will know it when you see it" seems to work more with class?

Ice, cannot disagree with your thinking, class is a slippery slope - anybody who says differently with backup would have my curiosity for sure. You added other parameters about weather and conditions which complicates things. I suppose in that case you should use races that are not under adverse condtion but sometimes you do not get that luxury. I had posted in another thread that for me at least until James Quinn book on Handicappers Condition Book came out no one was willing to define what class conditions were. For its time that was a bold undertaking and made sense at the time. LOL, some think he is still stuck in the past but is neither here nor there. Yes Quinn's book is now outdated but the bold original idea at least made a good attempt at trying to define at least in my mind why horses do not perform under certain race conditions and purse values when they look like they have the "goods" on paper and fail miserably against horses with less than stellar past performances. Back in the day race conditions were more uniform and consistent, now is it is so diluted with added conditions in order to fill the cards it is a joke at some venues IMO. All I am saying in broad brush terms that in general horses can only compete at certain levels once they have matured. If they deviate too much from that "range" they end up as an also ran. Like a 4yr old horse trying to run in an allowance race for the first time from the claiming ranks, not gonna see that horse in the winners circle very often and I can usually discount those types in my top tier selections. Best of luck.

[thaskalos]

To say class is one of a number of overlooked (or less emphasized in favor of more popular approaches) factors might be more descriptive.

There are approaches (a more appropriate term than specific strategies or tactics) that are as useful now as speed (and then pace) have been in the past. The advantage is in the fact that speed and pace still produce enough winners (with enough work and more than a little luck) to keep bettors chasing the same receding mutuel pools. Because bettors have become "comfortable" using the old approaches, they continue using them because they "kind of work." Some of the time. On a good day. When the moon is full. When those bettors contemplate change, a little voice whispers in their head, "Don't try to fix it--it ain't broke."

That little voice is lying, folks. It is broken--not because it doesn't occasionally generate a profit, and sometimes even an impressive profit--but because an equal amount of work applied in another area/other areas might be substantially more rewarding.

Consider how many times "alternative ideas" have been posted on this forum (besides the conventional approaches to speed and class) and those ideas have been immediately pooh-poohed by those claiming "generous ROIs every year, year after year." Or "seven-figure incomes from churning at the rebate shops", with no clue whatsoever how those doing so are (supposedly) generating a high enough rate of return WITHOUT the rebates to make the rebates "profitable" rather than just enticement for the big losers to keep losing big.

"Class"--as an approach to race analysis--is a topic that is complex enough, and difficult enough, to offer substantial advantage to those willing to do the work necessary.

I could do no more than just offer my own opinion...and here it is:

No matter what handicapping approach the player embraces...he will still only win "some of the time"...and he is guaranteed to spend considerable time scratching his head...after some of the race results he is bound to see.

I became a pace handicapper because I saw too many "speed" and "class" horses losing many races that they figured to win. But I then discovered that pace didn't offer the ultimate answer either. I am still surprised with some of the results I see...and it no longer bothers me. I now consider in an occupational hazard...

I use speed/pace not because I am more "comfortable" doing it; to be honest, it's a big pain in the ass for a "pencil-and-paper" handicapper like me.

I use speed/pace because no other approach has proven to me to be more profitable...and that's what is of utmost importance to me.

The sleepless nights I spend handicapping is the price that I have to pay, since, as we all know...nothing in life is free.

[traynor]

the issue I have with "class" is that it is hard to quantify and distribute that concept..Is their one uniform definition for it? Also, when a horse stepping up in class, it is clearly hard to determine if that horse belongs to this class or not.. until he wins or comes a close 2nd etc.. now take this by the large number of horses that are constantly going up or down and throw in some other factors like weather/track surface condition/bumps/steadying etc - which may cause a legitimate horse at a certain class level to still finish well below that level. How can class be determined and applied uniformly as a parameter?
"you will know it when you see it" seems to work more with class?

That is exactly what makes it so useful--it takes a bit more doing than writing a few lines of computer code and tweaking some numbers.

[traynor]

Ice, cannot disagree with your thinking, class is a slippery slope - anybody who says differently with backup would have my curiosity for sure. You added other parameters about weather and conditions which complicates things. I suppose in that case you should use races that are not under adverse condtion but sometimes you do not get that luxury. I had posted in another thread that for me at least until James Quinn book on Handicappers Condition Book came out no one was willing to define what class conditions were. For its time that was a bold undertaking and made sense at the time. LOL, some think he is still stuck in the past but is neither here nor there. Yes Quinn's book is now outdated but the bold original idea at least made a good attempt at trying to define at least in my mind why horses do not perform under certain race conditions and purse values when they look like they have the "goods" on paper and fail miserably against horses with less than stellar past performances. Back in the day race conditions were more uniform and consistent, now is it is so diluted with added conditions in order to fill the cards it is a joke at some venues IMO. All I am saying in broad brush terms that in general horses can only compete at certain levels once they have matured. If they deviate too much from that "range" they end up as an also ran. Like a 4yr old horse trying to run in an allowance race for the first time from the claiming ranks, not gonna see that horse in the winners circle very often and I can usually discount those types in my top tier selections. Best of luck.

The complexities introduced by all those added conditions are worth studying.

[traynor]

I could do no more than just offer my own opinion...and here it is:

No matter what handicapping approach the player embraces...he will still only win "some of the time"...and he is guaranteed to spend considerable time scratching his head...after some of the race results he is bound to see.

I became a pace handicapper because I saw too many "speed" and "class" horses losing many races that they figured to win. But I then discovered that pace didn't offer the ultimate answer either. I am still surprised with some of the results I see...and it no longer bothers me. I now consider in an occupational hazard...

I use speed/pace not because I am more "comfortable" doing it; to be honest, it's a big pain in the ass for a "pencil-and-paper" handicapper like me.

I use speed/pace because no other approach has proven to me to be more profitable...and that's what is of utmost importance to me.

The sleepless nights I spend handicapping is the price that I have to pay, since, as we all know...nothing in life is free.

It might be more appropriate to say that the too many speed and class horses you saw losing were defined as such by the criteria commonly used in those particular approaches at the time. Attaching a label to a horse as "class" based on simplistic criteria does not accurately reflect the class of the horse--it only reflects the application of certain criteria used to apply the label. If the crtieria used are deficient, the labels are spurious, and the usefulness of the labels diminished. It is not that the "class horse lost" as much as it is that the horse defined by people as being the "class horse" lost. The failure is in the people doing the labeling, not in the relative class of the horse.

It is always useful to consider Korzybski's admonition in Science and Sanity that the map is not the territory. Labeling a horse the "speed" horse in a race does not make it run faster.

I agree completely that there are no "magic bullets" that solve all problems and predict (accurately) the winner in every race. Handicapping is a process. Some of the newer processes are more predictive (and more profitable) than some of the older processes.

[Capper Al]

I can see that a good percentage of you are insisting in the concept of a 'class rating' expressed in numerical scale as a speed figure.

Although I do not believe such a metric can add any value to the bettor, I have to underline that such a rating is easy to to constructed (and most likely this is exactly why it adds no value).

The most important concept when creating a metric, is how to evaluate it. Once some handicapping attribute has been converted to a numeric quantity, its 'speciality' automatically becomes irrelevant. With this I mean that regardless of 'what' your figure is measuring it should be regarded as just a 'clear' (dimensionless) metric and nothing more. Such a metric can easily be compared to another and although in this posting I will not delve in to details conclusions can be drawn to decide which one is more effective. Following such an approach we can prove that the 'best' metrics are exactly those who replicate the behavior of the crowd; in other words if we were trying to create the ultimate figure having unlimited resources both in data and in computing power, chances are that we would end up with a composite encapsulating the fundamentals resulting to very high correlation with the crowd's odds. This is why effort consumed towards this direction has more academic importance than real betting value. As an example of this I refer you to this document:

http://eprints.usq.edu.au/4300/1/Williams_Li_Publ_version.pdf


Many years ago, when starting to apply pattern recognition and AI to race prediction I had discovered this document among other similar and replicated and improved their performance, only to find out that what I was doing was of no real value, as it would have been easier to reach the same conclusions just by eyeballing the board...

Fortunately, horse betting is not that simple. I say fortunately because otherwise everyone would follow an optimal approach quickly evaporating any possibility for profits!

If you really insist in class rating (or whatever else kind of metric) what you should concentrate is not the algorithm to use like for example:



as all these factors represent 'handicapping 101' material that will never give you any betting edge.

What might give you the edge, is to try to estimate what the probability and margin of error in the calculation of the figure.

This is the only estimate giving you some edge that you can transform to profits.

Needless to say that exactly the same applies not only to class but any other numerical figure you might be using....

You see nothing. I don't use the EPS or the APV or the Who beat Who method, but I do track the first two. Don't be overwhelmed with complexity. Simple can work and be more straight forward.

[thaskalos]

[QUOTE=traynor]It might be more appropriate to say that the too many speed and class horses you saw losing were defined as such by the criteria commonly used in those particular approaches at the time. Attaching a label to a horse as "class" based on simplistic criteria does not accurately reflect the class of the horse--it only reflects the application of certain criteria used to apply the label. If the crtieria used are deficient, the labels are spurious, and the usefulness of the labels diminished. It is not that the "class horse lost" as much as it is that the horse defined by people as being the "class horse" lost. The failure is in the people doing the labeling, not in the relative class of the horse.
/QUOTE]



No matter who does the "labeling", or what kind of method they use...there will still be a great difference between the horse that we see on paper, and the one that actually steps on the track.

And, if there are more profitable approaches out there than the ones that I am currently using...then I am confident that I will eventually stumble upon them.

After all...I am only 51.

[Capper Al]

AP 21 Jun 13
...........................................*** .. e-Ponies
Race Cond..............cls I.......cls II ......cls...... ML

#1 Clm 7500n3L ....3-4-2 ....2-7-4 ....2-1-4 ...2-4-7
#2 Clm 7500n2l .....7-4-2 ....2-4-5 ....3-4-6 ...5-2-7
#3 OClm 80000b ....2-6-5 ....2-6-4 .... 2-5-6 ...2-3-6

#4 OClm 40000n2x .2-4-3 ....4-7-6 ... 1-[5-4]* 4-7-[5-3-6]*
#5 Clm 25000n2l ....3-4-5 ....4-5-3 ... 5-3-2 ....5-2-4

#6 Md 15000 .........5-1-4 .....4-1-6 ...1-3-4 ....3-4-5
#7 Md Sp Wt .........2-4-1..... 2-5-4 .. 9-2-8 ....4-5-8

#8 Clm 5000b .......1-9-7 .....7-1-5....1-7-2 ....[1-1a]*-7-4
#9 Clm 25000b .....10-3-5 ...4-9-3 ... 6-9-5 ....9-10-6

[]* is tied

'cls I' is kind of like EPS while 'cls II' should be better because it considers form a little. Also included are e-Ponies class picks and the Morning Line for comparison.

Note races #6 and #7 are maidens but included anyway for discussion. These numbers are good as long as a first time starter doesn't win. I'm working on a way to handle this when I get time.

Please remember these are not final picks only class ratings for discussion.

Good Luck!

Not bad for no handicapping and it raining.

Hit
r1: #7 $8.20
r3: scr #2 -- #4 $7.60
r6: scr #4 -- #1 $8.80
r7: #2 $8.20

$2.00 dutch cost $36.00 returned $32.80 for a loss of $3.20.

[Maximillion]

Not bad for no handicapping and it raining.

Hit
r1: #7 $8.20
r3: scr #2 -- #4 $7.60
r6: scr #4 -- #1 $8.80
r7: #2 $8.20

$2.00 dutch cost $36.00 returned $32.80 for a loss of $3.20.


Im confused......not counting the morning line im just eyeballing the number of horses( "picked ?") in each race.Can you explain how you would then log these results in your records?

[traynor]

[QUOTE=traynor]It might be more appropriate to say that the too many speed and class horses you saw losing were defined as such by the criteria commonly used in those particular approaches at the time. Attaching a label to a horse as "class" based on simplistic criteria does not accurately reflect the class of the horse--it only reflects the application of certain criteria used to apply the label. If the crtieria used are deficient, the labels are spurious, and the usefulness of the labels diminished. It is not that the "class horse lost" as much as it is that the horse defined by people as being the "class horse" lost. The failure is in the people doing the labeling, not in the relative class of the horse.
/QUOTE]



No matter who does the "labeling", or what kind of method they use...there will still be a great difference between the horse that we see on paper, and the one that actually steps on the track.

And, if there are more profitable approaches out there than the ones that I am currently using...then I am confident that I will eventually stumble upon them.

After all...I am only 51.

Very true.

[Delta Cone]

Looking at two races at Hollywood Park on Saturday (6-22-2013) seems to demonstrate why the interplay of speed and class (and distance and surface) is so confusing.

Compare the conditions of race 1 with race 8. Both races are for 3yo and up, males

Race 1 (7f poly): Alw 54000n1x, purse of $54,000 and Beyer par of 94
Race 8 (8.5f turf): OC 62k/n2x, purse of $57,000 and Beyer par of 93

The Beyer pars suggest that these races are at the same level. Do you agree?

[raybo]

Continuing with the class figures we were posting yesterday. Al had suggested AP for today and Saturday, the track may be wet and off the turf but we'll see what happens anyway. I'm also continuing with Belmont today. Top 2 class rated horses unless there is a huge standout (1 will be posted), or a virtual tie in which case more than 2 will be posted. The ratings include a purse value metric, from competitive races only, a performance metric, and a form cycle/current form adjustment. Really pretty basic stuff, even the performance rating. Form cycles are 1-8 but only level 1 and 2 are considered for adjustments to the rating, also improvement/decline in form in the last 3 races add or subtract to/from the ratings (basically a "+", " -", or "=" notation is converted to +2, -2, or 0 points respectively). So, you can see this rating is not at all complex, compared to what some primarily class handicappers do. Don't know how an "off" track will affect these ratings so I will not address that scenario at all.

Also, no "'traditional" database research (unless you include past experience as a traditional database) was done regarding the creation of these ratings, just common sense. So, I guess the question is, how good is common sense in the creation of ratings, regardless of the type of ratings being created. Personally, I think common sense is of great importance in both handicapping methods and in wagering. Of course, any method that has been well designed is worthless if you have little patience, and/or discipline, and/or consistency in remaining true to the method.

APX 06/21/2013
Race 1 7.00f Dirt Clm 7500
2 91.6
1 83.5 -----2nd @ 2.1/1 odds
4 82.5
Race 2 9.00f Dirt Clm 7500
5 89.8 -----1st @ 2.4/1 odds ----$6.80
3 89.0
Race 3 -8.00f Turf OClm 80000clf
2 350.8 ----- scratched
Race 4 6.00f Dirt OClm 40000nw2
1 160.0 ------4th @ 5.4/1 odds
4 158.3
Race 5 -5.00f Turf Clm 25000
5 126.3 ------1st @ 1.6/1 odds ---$5.20
2 112.4 ------- 2nd @ 1.4/1 odds
Race 6 7.00f Dirt Md 15000
1 103.0 -------1st @ 3.4/1 odds ---$8.80
3 86.5 ---------4th @ 2.1/1 odds
Race 7 -5.00f Turf Md Sp Wt
1 118.0 ------scratched
8 114.5
Race 8 8.50f Dirt Clm 5000
1 136.2
4 108.1 -------2nd @ 4/1 odds
Race 9 -5.00f Turf Clm 25000
6 130.7 -------scratched
3 122.8
9 122.0 -------scratched


BEL 06/21/2013
Race 1 8.50f InnerT SOC
1 164.1 -------scratched
6 155.9
Race 2 5.00f Dirt Md Sp Wt
Pass - No data
Race 3 5.00f Dirt Md Sp Wt
Pass - No data
Race 4 8.50f Turf Md 35000
9 134.7 ---------scratched
4 118.3 1st @ 4/1 odds ---$10.00
Race 5 8.00f Dirt OClm 25000nw1
6 162.0 --------3rd @ 5.1/1 odds
1 155.0 --------4th @ 3.3/1 odds
Race 6 8.50f Turf Clm 20000
8 126.0 --------4th @ 14.5/1 odds
1 120.8 --------2nd @ 1.4/1 odds
9 120.0
Race 7 8.00f Dirt OClm 14000nw1
1A 158.0 --------3rd @ 1.7/1 odds
8 148.0 --------2nd @ 9.7/1 odds
1 147.0 --------1st @ 1.7/1 odds ---$5.40
Race 8 8.50f InnerT OClm 25000nw2
2 166.9
9 156.7 -------scratched
Race 9 8.50f Turf Md 20000
9 112.1 -------- 2nd @ 1.4/1 odds
8 110.3


Got laughed at today! Oh well, ain't the first time.

[Capper Al]

Im confused......not counting the morning line im just eyeballing the number of horses( "picked ?") in each race.Can you explain how you would then log these results in your records?

Use the 'cls II' column. I see now that the extra information is confusing to others. I like the extra info. For me, looking at how the trackman did on the same races gives me a better understanding on how my numbers did.

For tracking, these numbers are not used by themselves. They are part of a comprehensive formula that varies by race type. Look at e-Ponies for a similar approach that is explained on their website.

[Capper Al]

[QUOTE=traynor]It might be more appropriate to say that the too many speed and class horses you saw losing were defined as such by the criteria commonly used in those particular approaches at the time. Attaching a label to a horse as "class" based on simplistic criteria does not accurately reflect the class of the horse--it only reflects the application of certain criteria used to apply the label. If the crtieria used are deficient, the labels are spurious, and the usefulness of the labels diminished. It is not that the "class horse lost" as much as it is that the horse defined by people as being the "class horse" lost. The failure is in the people doing the labeling, not in the relative class of the horse.
/QUOTE]



No matter who does the "labeling", or what kind of method they use...there will still be a great difference between the horse that we see on paper, and the one that actually steps on the track.

And, if there are more profitable approaches out there than the ones that I am currently using...then I am confident that I will eventually stumble upon them.

After all...I am only 51.

Agree. The difference between what we figure and what is is enormous.

[traynor]

That may be why Beyer argued in The Winning Horseplayer almost 20 years ago that handicappers who relied exclusively on data--whether computer or pen-and-paper--were shooting themselves in the foot by using approaches with a plethora of built-in deficiencies. That many still believe numbers from Equibase are The Way and The Light can be attributed to good marketing by data providers and the continued support of that marketing by legions of software developers who rely on that data for income (via subscription fee commissions from the data providers). That may also be why any suggestion that the data available is insufficient are met with such hostility.

I go with the beliefs of the group from Argentina for which I developed an application that codified observations in a manner that could be integrated meaningfully with the available data. "Anyone who wagers on horse races should be close enough to the horses to smell them. No smell, no bet."

Fueling the myths of data supremacy are the auxiliary myths of the "whales" with super-sophisticated computer applications that do all the work, and enable them to make millions by punching a few keys on a computer console. Few suggest they might know more about horses and horse racing than the average bettor, and that superior knowledge is the primary component of their success--not a bigger database or a more complex software application.

People tend to believe what they want to believe, and to behave in a manner consistent with their beliefs. If one is using a deficient approach, it is difficult to admit that the approach is deficient, so fault is attributed to other factors.

[Robert Fischer]

Got laughed at today! Oh well, ain't the first time.
hey, that's part of the game.
I meant to say 'nice job on the Belmont card you posted the other day.
You came up pretty big in the last 3 races. :ThmbUp:

[JohnGalt1]

Looking at two races at Hollywood Park on Saturday (6-22-2013) seems to demonstrate why the interplay of speed and class (and distance and surface) is so confusing.

Compare the conditions of race 1 with race 8. Both races are for 3yo and up, males

Race 1 (7f poly): Alw 54000n1x, purse of $54,000 and Beyer par of 94
Race 8 (8.5f turf): OC 62k/n2x, purse of $57,000 and Beyer par of 93

The Beyer pars suggest that these races are at the same level. Do you agree?


My class comparison chart shows these two races are the same level, as is proven by the purse's similar amount.

[MJC922]

Looking at two races at Hollywood Park on Saturday (6-22-2013) seems to demonstrate why the interplay of speed and class (and distance and surface) is so confusing.

Compare the conditions of race 1 with race 8. Both races are for 3yo and up, males

Race 1 (7f poly): Alw 54000n1x, purse of $54,000 and Beyer par of 94
Race 8 (8.5f turf): OC 62k/n2x, purse of $57,000 and Beyer par of 93

The Beyer pars suggest that these races are at the same level. Do you agree?

Though I don't use pars of any kind to calculate class ratings, I do sometimes include a class par in my product for eyeballing purposes if the sample size is large enough at that specific track. For these two races I happen to have the first race par of 11 and the eighth race is 8.75. Because my numbers are on a different scale (a ragozin style scale) the eighth race would be a solid notch better. Not sure why the Beyer par is close and indicates the opposite. Is that par printed in the DRF?

[traynor]

The simplest way to explain it may be that I am very much in favor of a research process called "triangulation" applied to horse racing. Specifically, triangulation is a synergistic combination of quantitative (database/data centric/number crunching) analysis and qualitative (trips/body language/paddock inspection/critically watching races) analysis. The combination should be complimentary, and I think the success of the majority of "whales" (or those referred to as whales who are not simply large losers funded from sources outside of racing) is directly attributable to such methods.

If profit is the primary motive, either approach--used exclusively--is an error.

[Capper Al]

hey, that's part of the game.
I meant to say 'nice job on the Belmont card you posted the other day.
You came up pretty big in the last 3 races. :ThmbUp:

You're right. Arlington had a lot of turf to dirt yesterday. I appreciate bloggers who post their ideas ahead of the races, not just pie in the theory. You killed them at Belmont.

[Capper Al]

The simplest way to explain it may be that I am very much in favor of a research process called "triangulation" applied to horse racing. Specifically, triangulation is a synergistic combination of quantitative (database/data centric/number crunching) analysis and qualitative (trips/body language/paddock inspection/critically watching races) analysis. The combination should be complimentary, and I think the success of the majority of "whales" (or those referred to as whales who are not simply large losers funded from sources outside of racing) is directly attributable to such methods.

If profit is the primary motive, either approach--used exclusively--is an error.

Agree. If I was in a syndicate, such as a profitable whale might be, I'd expect to have a trainer or vet on my team for the physical non-numeric evaluation. We'd only make a wager when both teams agreed.

[DeltaLover]

The simplest way to explain it may be that I am very much in favor of a research process called "triangulation" applied to horse racing. Specifically, triangulation is a synergistic combination of quantitative (database/data centric/number crunching) analysis and qualitative (trips/body language/paddock inspection/critically watching races) analysis. The combination should be complimentary, and I think the success of the majority of "whales" (or those referred to as whales who are not simply large losers funded from sources outside of racing) is directly attributable to such methods.

If profit is the primary motive, either approach--used exclusively--is an error.

I maintain a different view, adopting a very contrarian approach to what you describe here.

Your approach although quite reasonable for pure 'handicapping' purposes seems to favor determinism placing stochasticity to a second class citizen as far as the final betting decision goes.

For betting purposes I am convinced that the later is way more significant than the former.

I think that the most important component of a successful betting strategy is to understand the crowd and look for cases where is is terribly wrong.

I know, this statement does not bring anything new to our conversations... We have been talking about it for long time...

Still though, I am under the impression that very few are deeply convinced about its validity and still most of us we are looking for magic solutions that will put us ahead of the crowd with things like body language or paddock inspection.

This approach does not put us ahead of the curve, at least not to the point to show betting profits. The reason is that since body language or paddock inspection (or anything similar) do not represent major sources of the crowd's opinion and more important can not be objectively defined makes it very difficult to search for betting indecencies based on them!

I realize of course that a factor that is completely unknown to the public has the potential to identify its errors but this will happen only in extreme cases, body language or paddock inspection not been two of them.

An example of this can be found in beyer speed figures. They probably were money making tools prior to their adoption from DRF. But this was happening because they represent a very strong indicator of the final outcome of the race allowing its derivatives (like bris prime power) to show extremely high impact values.

I really do not believe you can come up with any kind of a rating even remotely close to this level of accuracy by inspecting body language. If you excel in it you might be able to find a slim advantage and even then you will need to put a lot of work on it, keeping detailed records, constantly watching post parades and warm ups etc.

[MJC922]

The simplest way to explain it may be that I am very much in favor of a research process called "triangulation" applied to horse racing. Specifically, triangulation is a synergistic combination of quantitative (database/data centric/number crunching) analysis and qualitative (trips/body language/paddock inspection/critically watching races) analysis. The combination should be complimentary, and I think the success of the majority of "whales" (or those referred to as whales who are not simply large losers funded from sources outside of racing) is directly attributable to such methods.

If profit is the primary motive, either approach--used exclusively--is an error.

Couldn't agree with you more on all of the above.

[Robert Fischer]

Looking at two races at Hollywood Park on Saturday (6-22-2013) seems to demonstrate why the interplay of speed and class (and distance and surface) is so confusing.

Compare the conditions of race 1 with race 8. Both races are for 3yo and up, males

Race 1 (7f poly): Alw 54000n1x, purse of $54,000 and Beyer par of 94
Race 8 (8.5f turf): OC 62k/n2x, purse of $57,000 and Beyer par of 93

The Beyer pars suggest that these races are at the same level. Do you agree?

Here is what I have.


Race 1 looks to be the classier race.
:3: Everton figures to run well at a short price. :1: Motown Men looks competitive ml 6-1.


Race 8 may be a good betting race. important to note: *:6: King Edward comes off 2yr Layoff and can't be expected to run to his class number. :5: He Be Fire N Ice gives the race some betting value, as he is a weak favorite and can be bet against. :2: Ivanho draws a soft spot here, and picks up Stevens at a square price. :7: Kingpin Ryno 2nd off the 10mth layoff, may have needed that sprint last time and gets some pace relief here, the kind of clunker who isn't much worse and could go off higher than his ml odds.

HOL 06/22/2013

Race 1 7f ALW54n1x
:3: 95
:1: 89
:4: 88

Race 8 8.5f Turf OC62.5n2x
:6: 93*
:2: 89
:1: 89
:5: 88
:7: 86

[traynor]

Agree. If I was in a syndicate, such as a profitable whale might be, I'd expect to have a trainer or vet on my team for the physical non-numeric evaluation. We'd only make a wager when both teams agreed.

That is really close. Multiple, complimentary perspectives are often much more accurate than the isolated perspectives. It is not essential that the perspectives be provided by different people--I know several very serious bettors who take their computer printout (or mobile app with the output) to the track, and make final decisions based on multiple perspectives.

The really odd thing is that in 20 years or so of associating with a wide variety of bettors that most would consider "very serious" NONE--not even one--has relied exclusively on computer output for wagering. Because I meet most onsite at the track, that may be a skewed perception. However, the individuals and groups I have encountered in various colleges and universities (where I also spend a lot of time) had similar inclinations. While most were continually refining and improving their computer apps, none considered those apps as the be all-end all of race analysis.

In that 20 years or so of hobnobbing with serious bettors and betting groups, I have never encountered anyone, nor have I ever encountered anyone who had ever encountered anyone, who believed they could sit at a computer crunching numbers, betting online, and making huge profits as if playing a videogame. Or by "churning" wagers for rebates.

[traynor]

I maintain a different view, adopting a very contrarian approach to what you describe here.

Your approach although quite reasonable for pure 'handicapping' purposes seems to favor determinism placing stochasticity to a second class citizen as far as the final betting decision goes.

For betting purposes I am convinced that the later is way more significant than the former.

I think that the most important component of a successful betting strategy is to understand the crowd and look for cases where is is terribly wrong.

I know, this statement does not bring anything new to our conversations... We have been talking about it for long time...

Still though, I am under the impression that very few are deeply convinced about its validity and still most of us we are looking for magic solutions that will put us ahead of the crowd with things like body language or paddock inspection.

This approach does not put us ahead of the curve, at least not to the point to show betting profits. The reason is that since body language or paddock inspection (or anything similar) do not represent major sources of the crowd's opinion and more important can not be objectively defined makes it very difficult to search for betting indecencies based on them!

I realize of course that a factor that is completely unknown to the public has the potential to identify its errors but this will happen only in extreme cases, body language or paddock inspection not been two of them.

An example of this can be found in beyer speed figures. They probably were money making tools prior to their adoption from DRF. But this was happening because they represent a very strong indicator of the final outcome of the race allowing its derivatives (like bris prime power) to show extremely high impact values.

I really do not believe you can come up with any kind of a rating even remotely close to this level of accuracy by inspecting body language. If you excel in it you might be able to find a slim advantage and even then you will need to put a lot of work on it, keeping detailed records, constantly watching post parades and warm ups etc.

Yes, but not for simple reasons. Stochastic processes work well in some situations. Horse racing is not one of those situations.

[MJC922]

The reason is that since body language or paddock inspection (or anything similar) do not represent major sources of the crowd's opinion and more important can not be objectively defined makes it very difficult to search for betting indecencies based on them!


I disagree, not with all of your post as you're obviously astute and make very good points through the course of it, however the above slice I do disagree with in the sense that you're implying somehow which factors the crowd looks at (or in this case doesn't look at) has something to do with probabilities. Because a factor is profitable or not or is reflected in large part on the tote has nothing to do with how predictive it is, rather it has to do with how well distributed it is, I would not want to confuse the two. These physical characteristics of horses can lead to situations which open up the door to killing specific races as there are times when only two horses can be played and one can barely walk up to the gate. I think when you look back on a profitable meet these types of scenarios can be a significant source (or even the entire source) of your profits. That it's very difficult to quantify I can't argue with but it's the one relevant area left largely unexplored. Pittsburg Phil was convinced physical observation was the key to his success, and when I read the Maxims (in my view the best handicapping book ever written by a wide margin) in a mere page or two of his second chapter he describes his mode of operation which quite frankly makes modern day players (including me) look totally inadequate by comparison.

[DeltaLover]

Yes, but not for simple reasons. Stochastic processes work well in some situations. Horse racing is not one of those situations.

You mean that you always approach your betting selection using a pure deterministic decission making process?

[traynor]

"In science (/wiki/Science), the term observer effect refers to changes that the act of observation will make on a phenomenon (/wiki/Phenomenon) being observed. This is often the result of instruments that, by necessity, alter the state of what they measure in some manner. A commonplace example is checking the pressure in an automobile tire; this is difficult to do without letting out some of the air, thus changing the pressure. This effect can be observed in many domains of physics.

The observer effect on a physical process can often be reduced to insignificance by using better instruments or observation techniques. However in quantum mechanics (/wiki/Quantum_mechanics), which deals with very small objects, it is not possible to observe a system without changing the system, so the observer must be considered part of the system being observed."
http://en.wikipedia.org/wiki/Observer_effect_(physics)

Many consider pari-mutuel racing an interactive process, in which knowledge of the data changes the data. In simplistic terms, the data (as it existed) depended on a lack of knowledge of its existence. Once that data is observed and becomes known, it will not be replicated.

That is the underlying reason that regression analysis and database handicapping suffer from shrinking profits--the reason the profits existed in the first place is because no one knew they existed. That is the underlying reason that data models based on races in which wagers were not made (by the researcher) are less accurate than models based on races in which wagers of a given amount were actually made.

In short, stuff always looks better on paper than it does in the real world.

[traynor]

You mean that you always approach your betting selection using a pure deterministic decission making process?

Not at all. As I have described previously, I make a large number of (relatively small) wagers using purely mechanical selection processes from a computer app. When I wager serious amounts (very limited number of races, very limited number of tracks, usually one, sometimes two or three, but highly selective) I usually do so at the track, for which I have extensive trip notes and other information to compliment the computer output. I don't make "serious" wagers based solely on computer output. I want to see the horses up close before I make any final decisions.

Do I think I am competent enough at observational information processing to make a difference? Yes.

[DeltaLover]

"In science (/wiki/Science), the term observer effect refers to changes that the act of observation will make on a phenomenon (/wiki/Phenomenon) being observed. This is often the result of instruments that, by necessity, alter the state of what they measure in some manner. A commonplace example is checking the pressure in an automobile tire; this is difficult to do without letting out some of the air, thus changing the pressure. This effect can be observed in many domains of physics.

The observer effect on a physical process can often be reduced to insignificance by using better instruments or observation techniques. However in quantum mechanics (/wiki/Quantum_mechanics), which deals with very small objects, it is not possible to observe a system without changing the system, so the observer must be considered part of the system being observed."
http://en.wikipedia.org/wiki/Observer_effect_(physics)

Many consider pari-mutuel racing an interactive process, in which knowledge of the data changes the data. In simplistic terms, the data (as it existed) depended on a lack of knowledge of its existence. Once that data is observed and becomes known, it will not be replicated.

That is the underlying reason that regression analysis and database handicapping suffer from shrinking profits--the reason the profits existed in the first place is because no one knew they existed. That is the underlying reason that data models based on races in which wagers were not made (by the researcher) are less accurate than models based on races in which wagers of a given amount were actually made.

In short, stuff always looks better on paper than it does in the real world.


Nice posting!

I have to add that introducing a meta handicapping level, where a handicapping approach is handicapped can potentialy solve the problem of shrinking profits.

Of course, this assumes crowd's asymptotic approach to the real probability distribution; if this does hold true the game can become unprofitable.

[DeltaLover]

Do I think I am competent enough at observational information processing to make a difference? Yes.

Good for you Mr Traynor...

I cannot make a similar claim though...

[Delta Cone]

Though I don't use pars of any kind to calculate class ratings, I do sometimes include a class par in my product for eyeballing purposes if the sample size is large enough at that specific track. For these two races I happen to have the first race par of 11 and the eighth race is 8.75. Because my numbers are on a different scale (a ragozin style scale) the eighth race would be a solid notch better. Not sure why the Beyer par is close and indicates the opposite. Is that par printed in the DRF?

Yes, the Beyer par is printed in the DRF. However, I don't know what they are using to calculate the par (i.e., how many races, do they separate out sprint n1x from route n1x?, etc.)

[Delta Cone]

Here is what I have.


Race 1 looks to be the classier race.
:3: Everton figures to run well at a short price. :1: Motown Men looks competitive ml 6-1.


Race 8 may be a good betting race. important to note: *:6: King Edward comes off 2yr Layoff and can't be expected to run to his class number. :5: He Be Fire N Ice gives the race some betting value, as he is a weak favorite and can be bet against. :2: Ivanho draws a soft spot here, and picks up Stevens at a square price. :7: Kingpin Ryno 2nd off the 10mth layoff, may have needed that sprint last time and gets some pace relief here, the kind of clunker who isn't much worse and could go off higher than his ml odds.

HOL 06/22/2013

Race 1 7f ALW54n1x
:3: 95
:1: 89
:4: 88

Race 8 8.5f Turf OC62.5n2x
:6: 93*
:2: 89
:1: 89
:5: 88
:7: 86

Thanks for the analysis. I think that you have shown that the class conditions and Beyer pars don't really tell you the whole story about the class level of that specific race. One needs to look at the entrants that actually run.

This leads to problems when looking at PPs months later. "Now, let me see, was this n1x a strong one, or weak for the level?" I currently have no way to measure this.

[lsosa54]

Yes, the Beyer par is printed in the DRF. However, I don't know what they are using to calculate the par (i.e., how many races, do they separate out sprint n1x from route n1x?, etc.)

Having handicapped thousands of both types of races, mainly on the SoCal circuit, they are not the same. Could a horse from the N1X race win the optional claimer, sure. I would prefer he have won his n1x condition strongly first.

No matter the speed figs or pars, my records show the preference to be a lightly raced horse that has run well or shown life at this condit1on previously. A very sharp claimer at the 50-62.5 level can also be formidable, sometimes even a very sharp winner at the 40K level.

My last preference would be the n1x horse in the optional claimer, unless he had won the n1x and was a nicely bred horse just cruising through his non claiming conditions.

For me it also varies by time of year, giving the typical n1x horse a better chance to win the optional claimer later in the year, when most of the good horses have cleared the condition. The "optinal claiming" part of thr condition really does change the situation.

As for your comment right above, you could record the adjusted fractional and final times for each race to see what races may be stronger. Only problem with that in isolation is that you probably also need to record the running styles of the race entries to see how that affected the times.

[PICSIX]

Churchill Downs 6/22

[Robert Fischer]

another issue with comparing the 2 races at Hollywood, is the surface/distance requirements.

Everton is the best horse in either race, but he'd be less of a lock in the 8.5f turf race. More because of the requirements of that race, and his light résumé, than anything else.

Both groups in general are probably better placed where they are at, with the exception of maybe Circle the Moon in the 8th, who could be the type to appreciate 7f on cushion/main track.