Published Speed Figures

[douglasw32]

Help me out... if you were going to purchase speed figures what provider ?

I would like to keep it as cheap as possible, please give me your 2 cents on who is better than who out there when it comes to just getting the speed figure right ?

[completebill]

While I haven't purchased it in years, my previous experience was that the adjusted fractional and final times published in Todays Racing Digest ere EXTREMELY good.

I don't know whether the figures generated by Jim Cramer at Handicappers Data Warehouse are available as a stand-alone product, but they are just SUPERIOR. They are included in the data supplied to several of the major handicapping programs, including HTR and HSH

[riskman]

While I haven't purchased it in years, my previous experience was that the adjusted fractional and final times published in Todays Racing Digest ere EXTREMELY good.

I don't know whether the figures generated by Jim Cramer at Handicappers Data Warehouse are available as a stand-alone product, but they are just SUPERIOR. They are included in the data supplied to several of the major handicapping programs, including HTR and HSH

I use HDW figures and have for a long time.

[46zilzal]

I get data from one of these no speed figures

[Skanoochies]

Why isn`t C.J. on the list. I hear he is as good as any? :confused:

[Cardus]

Why isn`t C.J. on the list. I hear he is as good as any? :confused:

Echo.

[douglasw32]

Why isn`t C.J. on the list. I hear he is as good as any? :confused:

I thought the BUY in was limited ?

[Fingal]

To use as a stand alone number after selecting the contenders, I'd go for Today's Racing Digest followed by BRIS.

[exiles]

[QUOTE=douglasw32]Help me out... if you were going to purchase speed figures what provider ?

I would like to keep it as cheap as possible, please give me your 2 cents on who is better than who out there when it comes to just getting the speed figure right ?[/QUOTE
T

IMO The RAGOZINS have no equal but they are very expensive. As for the worst BRIS by a lot.

[bisket]

take one piece of advise from da bisket. before you begin to use any speed figures make sure to fully acquaint yourself with the formula used to arrive at a figure. if the formula isn't provided don't use it. to properly use speed figures in your own handicapping you need to make sure you aren't taking something into consideration that the figure has already accounted for. you also may have a differing opinion than the figure maker on that one particular point or not agree with weight the figure maker gives this particular piece of information. the figure maker could consider one length behind the pace leader a stalker and you feel a 1/2 length denotes a stalker.....etc. you have to know how a figure is made to properly put it into your routine.

[Tom]

CJ is number one, but, like you say, limited distribution. The best, and the cheapest. Go figure!

Frankly, if you don't want to spend a C note or better a month for Crammer figures, the Beyers are the best of the rest. Speed Figures.com is limited to California, but very good Quirin style numbers.

[Robert Goren]

At one time Kruk was as good as it got. The last few years John put them out they were a little lacking. CJ's is about as good as they get these days. I do not know anything about Cramer's. Rogozin is not really what I would call a speed figure like Beyers. They all have problems, some more than others. The main gripe I have is that you know how much faith to put a single races and or days rating. The rating makers would improve their product if they rated rating. Say a 5 for we think this number pretty close to right to a 1 for we have a lot of doubts about this number. All that said Beyers are pretty good at the big league tracks especially on races ran on a Saturday. For small tracks, I don't trust antibody's very far. JMO

[Gapfire]

CJ is number one, but, like you say, limited distribution. The best, and the cheapest. Go figure!

Frankly, if you don't want to spend a C note or better a month for Crammer figures, the Beyers are the best of the rest. Speed Figures.com is limited to California, but very good Quirin style numbers.

With all due respect, I think gapfire's are just as good, and we are priced more competitively than the others.

[Gapfire]

I think most of these services put out good figures. It would also be very hard to say who the best is unless you played a ton of races with all of them.

[Capper Al]

I adjust BRIS speed figures. My research has shown that most of these vary a lot from race to race but tend to overall win at the same rate in the long run. The trick might be to know whose speed figure to use for a race type rather than commit to one single source overall.

[douglasw32]

With all due respect, I think gapfire's are just as good, and we are priced more competitively than the others.

Just an FYI:

The quickgap is working but on your site , tons of sql errors in the advanced section for Belmont today.

And many sql errors even with the quickgap as you move through the tracks.

[fmolf]

I have found the bris speed and pace figs to be accurate enough for the way i handicap and their is loads of other good data and ratings and info included in their ultimate pp's....i have also used trackmasters top product platinum i believe they call them and like the way they rank the horses against each other in the race at hand and also like the way the organize their trainer and jockey info......I in no way would be considered a speed handicapper and use figs as only one aspect of my handicapping.

[lamboguy]

the best speed figures are greg matties numbers. the reason why is because he has ontrack input as to where the bias on the track is for a particular race card. that being said, they don't sell their figures and they are not looking for new people to participate in their program.

as far as cj's figures go they are very good as well. the way they are presented is pretty neat with the quirrin points. he breaks the race down to 2 different segments early and late pace. i originally thought early and late wasn't that good, but when i figured it out it was alot more easier to comprehend than the guys that break it down into thirds. he has a bias line that is based on the total amount of races during the day for the total race whether speed won or lost. he does not take into consideration inside and outside bias's just total bias.

the real bottom line is there are going to be flaws with every system of pace ananlysis. for the money cj's are well worth it because i seem to find them very accurate for dirt races and a distance variences like 5 1/2 fulongs to 6. i have found that it is impossible to rely on any speed figure for 4 1/2 furlong races.

if you are a bigtime turf player i would suggest ragozin or therograph, but those are very expensive and since i stay away from turf i don't use them.

by the way i ran into a circumstance where the buyer number came out at 46 and that race got upgraded to a 50 2 months later without the horse running. cj had the number right immediately.

i personally use something else that rates the probability of a horse winning the race based on pace. it is very subjective and it has a problem of rating horses from one distance to the next and one track from the other. but for me it helps me to eliminate races that i play where they are impossible to figure. also they don't do every track like cj.

[Gapfire]

Just an FYI:

The quickgap is working but on your site , tons of sql errors in the advanced section for Belmont today.

And many sql errors even with the quickgap as you move through the tracks.

I see you bought a day pass, thank you. The Quickgaps for Belmont are fine. You just caught them as they were calculating.

There will also be errors at Turfway today, because they are using KY downs as part of their card. I haven't got the Quickgaps set up to calculate that track.

If the track is on our list, and the track is running today, there will be Quickgaps.

Thank you,

Kevin
www.gapfire.com

[Fingal]

Speed Figures.com is limited to California, but very good Quirin style numbers.

One caveat about them, while the website says they post them before 5 PM PST, many times the S Cal ones aren't. That makes it kinda rough on someone that likes to have all the grunt work done the evening before & uses the mornings to go back over each race & fine tune.

They are very good numbers though, especially when one takes into account the track profile.

[Gapfire]

Just an FYI:

The quickgap is working but on your site , tons of sql errors in the advanced section for Belmont today.

And many sql errors even with the quickgap as you move through the tracks.

Also, the Quickgaps for Belmont will be updated for scratches when they come out. You may have to refresh your browser to get the latest Quickgaps.

Kevin

www.gapfire.com

[douglasw32]

I see you bought a day pass, thank you. The Quickgaps for Belmont are fine. You just caught them as they were calculating.

There will also be errors at Turfway today, because they are using KY downs as part of their card. I haven't got the Quickgaps set up to calculate that track.

If the track is on our list, and the track is running today, there will be Quickgaps.

Thank you,

Kevin
www.gapfire.com (http://www.gapfire.com)

Ahhhh, Thanks Kevin ;)

[Tom]

Gapfire....didn't mean to slight you. I was only addressing the others on the poll - didn't know you made them, too.

[Gapfire]

Gapfire....didn't mean to slight you. I was only addressing the others on the poll - didn't know you made them, too.

No offense taken. It just gave me an opportunity to let people know about our figures.

[Greyfox]

With all due respect, I think gapfire's are just as good, and we are priced more competitively than the others.

Thanks Kevin.
Your Dad's comments in the "Gapfire Pick Area" made me re-evaluate my tri play in Race 12. Got it.
Longshot # 5 Oly's Offspring just made third. I'd originally tossed him.

[classhandicapper]

take one piece of advise from da bisket. before you begin to use any speed figures make sure to fully acquaint yourself with the formula used to arrive at a figure. if the formula isn't provided don't use it. to properly use speed figures in your own handicapping you need to make sure you aren't taking something into consideration that the figure has already accounted for. you also may have a differing opinion than the figure maker on that one particular point or not agree with weight the figure maker gives this particular piece of information. the figure maker could consider one length behind the pace leader a stalker and you feel a 1/2 length denotes a stalker.....etc. you have to know how a figure is made to properly put it into your routine.

I totally agree with this and then some.

I've given countless hours of thought into trying to understand the pluses and minuses of the various approaches used by the leading figure makers. There are clear cut differences in methodology that account for the different figures given to the same horse by different figure makers for the same race. If you don't understand them, you are screwed.

If I can find the time and some of my youthful motivation again, I may start making comparative pace/finish class figures for NY and some of the major stakes around the country.

[thaskalos]

you also may have a differing opinion than the figure maker on that one particular point or not agree with weight the figure maker gives this particular piece of information. the figure maker could consider one length behind the pace leader a stalker and you feel a 1/2 length denotes a stalker.....etc. you have to know how a figure is made to properly put it into your routine.I thought that speed figures only told us how fast a horse completed a race, bisket.

What do things like "presser" and "stalker" labels have to do with the speed figures themselves?

The speed figures are the raw tools. The analysis that we do with these figures, falls under the category of "interpretation of the figures"...which is an altogether different, and more important step, IMO.

[cj's dad]

WWW,pacefigures.com

The best but then I am prejudiced.

[classhandicapper]

WWW,pacefigures.com

The best but then I am prejudiced.

Yes, they are quite good aren't they. ;)

[Fastracehorse]

I never get tired of them; in fact I love 'em but I already said that.

Players who don't like them probably read them too literally.

There is a bold type # printed for you in the form called the Beyer - and you have to ADJUST it based on the context of the game - and that is entirely subjective because approaches towards selecting a choice vary greatly - the approach is also contextual.

They can and do pick longshots: not the only way - but when available they help to eliminate alot of confusion; which bounds aplenty in our game.

fffastt

[protop]

without the speed fig;s you are betting blind

[thaskalos]

without the speed fig;s you are betting blindMaybe...but placing too much emphasis on them is like betting with only one eye open.

[PhantomOnTour]

Why is it that folks think all fig makers/users just up and bet the top last race fig? Almost every single non fig handicapper I encounter has that illusion that we are just rote betters playing the highest figure....rubbish!

[cj's dad]

Why is it that folks think all fig makers/users just up and bet the top last race fig? Almost every single non fig handicapper I encounter has that illusion that we are just rote betters playing the highest figure....rubbish!

You nailed it. When looking at the figs I use several factors come into play as well as the info available in the DRF - they go hand in hand IMO

[Tom]

Those most vocal against figs really have no clue how we used them.

[Charlie D]

Those most vocal against figs really have no clue how we used them.

I would tend to agree Tom, however, would you not agree that anyone who thinks the below is probably a blind person??

without the speed fig;s you are betting blind

[cj's dad]

I would tend to agree Tom, however, would you not agree that anyone who thinks the below is probably a blind person??

without the speed fig;s you are betting blindhe equation


No, but I believe one is missing a sizeable part of the equation !

[thaskalos]

There is one thing I don't understand about speed figures. Why is there such a discrepency between different figure makers...as far as their day-to-day figures are concerned.

I understand that the most advanced figure makers employ "projected" variants...which allow human intervention into the figure making process...but why is there such a difference among the computer generated, par-time based, speed figures?

Both the Bris figures and the Equibase figures are computer generated...and yet...the inconsistency between them is striking.

[PhantomOnTour]

There is one thing I don't understand about speed figures. Why is there such a discrepency between different figure makers...as far as their day-to-day figures are concerned.

I understand that the most advanced figure makers employ "projected" variants...which allow human intervention into the figure making process...but why is there such a difference among the computer generated, par-time based, speed figures?

Both the Bris figures and the Equibase figures are computer generated...and yet...the inconsistency between them is striking.
A few reasons for discrepancies:
1) fig makers put a different numerical value on a beaten length and 1/5 second. Quirin style says 1 length=1/5sec=1pt for final time and 1 length=1/10sec=1pt for pace. Beyer (who doesn't do pace) has a different value for a beaten length at every distance. Throw in Rags,BRIS,Equibase etc...and you have many different ways to arrive at a 'number'.
2) some days require serious work to arrive at a track variant (which is the rate at which you will adjust the actual running times, and the key to the whole equation imo). The TV purports to tell how fast or slow the surface was in relation to a par or projected time. Some figgers simply interpret the day's data differently; either omitting a race from the variant calculation, splitting the variant due to extreme weather changes during the card, re-rating a race with a runaway winner in fantastic time, you name it.....
Computer gen numbers probably vary because of the purely human decision making that goes into making the TV for some days, and these computers can't do it. Just a guess as I make all mine.

If you asked the fig makers of this board to post their variant for a single day of racing (at any track they cover) you would probably find our numbers to be close, but by no means exactly the same.
In fact, I think it's neat experiment to do. We could choose a day from earlier this year so folks don't give up the goods, so to speak, and compare our differences.

[Charlie D]

without the speed fig;s you are betting blindhe equation


No, but I believe one is missing a sizeable part of the equation !


You are well within your rights to believe that cj dad, but as a PA poster once correctly stated . "There is more than one way" to make the equation.

[mistergee]

has anyone here tried to use the beyer "pace" figures that he explained how to achieve in one of his books? How did they do, and how to you apply them in conjuction to final figures? Thanks

[Fastracehorse]

Maybe...but placing too much emphasis on them is like betting with only one eye open.

.......is the approach to sound mgmt of figs?

I've said this before but speed figs can teach players about the game - such as how trainers manage the form of the horse over a span of 4 races. The game is surprisingly universal - or not so surprising because it has a deep history - and alot of training methods from the 1930's are relevant today.

Also, knowing figs weakness is a knowledge of it's own; hence, furthering the efficacy of the decision making preocess.

Lot to learn,

fffastt

[Fastracehorse]

I would tend to agree Tom, however, would you not agree that anyone who thinks the below is probably a blind person??

.......of vision impairment when U can't tell how good a horse is. Yah, class to class comparison is good; key races, and trainer intent, etc; are all part of the game but the fig just makes everything crystal clear.

I can say this cuz i used to bet without them.

fffastt

[thaskalos]

.......is the approach to sound mgmt of figs?

I've said this before but speed figs can teach players about the game - such as how trainers manage the form of the horse over a span of 4 races. The game is surprisingly universal - or not so surprising because it has a deep history - and alot of training methods from the 1930's are relevant today.

Also, knowing figs weakness is a knowledge of it's own; hence, furthering the efficacy of the decision making preocess.

Lot to learn,

fffasttWHO SAID TOO MUCH EMPHASIS?

I was responding to protoc, who stated that : "Without the speed figs, you are betting blind."

Thanks for the lesson though...I'll try to keep it in mind...

[Charlie D]

.......of vision impairment when U can't tell how good a horse is. Yah, class to class comparison is good; key races, and trainer intent, etc; are all part of the game but the fig just makes everything crystal clear.

I can say this cuz i used to bet without them.

fffastt


Wasn't me backing Rachel at odds on Fast, wasn't me betting Vineyard Haven, as deafeat was out of the question . So i guess it must have been those people with the figs that make everything crystal clear.

[TrifectaMike]

There is one thing I don't understand about speed figures. Why is there such a discrepency between different figure makers...as far as their day-to-day figures are concerned.

I understand that the most advanced figure makers employ "projected" variants...which allow human intervention into the figure making process...but why is there such a difference among the computer generated, par-time based, speed figures?

Both the Bris figures and the Equibase figures are computer generated...and yet...the inconsistency between them is striking.

Aside from the methodology used to derive the figures, they are presented as a number, when it should be a distribution. All speed figures are subject to measurement errors, some more than others.

A speed figure is comprised of
estimated speed figure = true speed figure + error

Only God knows the true speed figure, and the error has two components: random noise and systematic.

The random error affects the variability of the estimated speed rating. Whereas the systematic error affects the mean vaule of the estimated speed rating. This results in a bias.

One of the best things you can do to deal with measurement errors, especially systematic errors, is to use multiple measures of the same estimated speed figure. Especially if the different measures don't share the same systematic errors, you will be able to triangulate across the multiple estimates and get a more accurate sense of what's going on.

Mike

[cj]

Aside from the methodology used to derive the figures, they are presented as a number, when it should be a distribution. All speed figures are subject to measurement errors, some more than others.

A speed figure is comprised of
estimated speed figure = true speed figure + error

Only God knows the true speed figure, and the error has two components: random noise and systematic.

The random error affects the variability of the estimated speed rating. Whereas the systematic error affects the mean vaule of the estimated speed rating. This results in a bias.

One of the best things you can do to deal with measurement errors, especially systematic errors, is to use multiple measures of the same estimated speed figure. Especially if the different measures don't share the same systematic errors, you will be able to triangulate across the multiple estimates and get a more accurate sense of what's going on.

Mike

I agree with this, but I'm not sure the reception figures would get if you said Ghostzapper ran 128+5 and Quality Road 121+7. It would definitely be something I would recommend to a person making figures for themselves though.

[douglasw32]

Wait dumb this one down for me guys =)

Do you mean look at say Bris, equibase and DRF figures and come up with how they all relate.

[TrifectaMike]

Wait dumb this one down for me guys =)

Do you mean look at say Bris, equibase and DRF figures and come up with how they all relate.

Correct me if I'm wrong. I believe you are referring to the term, Triangulation.
If that is the case, you are correct about using various sources as well as different methods to estimate a speed rating.

The value of triangulation is to determine if convergence exists. Using the data sources (Bris, Equibase, and DRf figures), we would like to see evidence of the convergence of ratings. Sometimes convergence occurs, however, it is also likely it might not.

In fact, there are three possible outcomes that might result from a Triangulation strategy.

The first is the goal you are seeking convergence. That is agreement amongst the different sources or methods.

The second is inconsistency. That is the ratings may be inconsistent, not confirming but not contradictory.

The third is contradictory. The estimated ratings from differing methods and sources contradict each other.

In each case the estimated speed ratings have to be filtered through knowledge gleaned from the data, context, and understanding of the process.

I hope this answers your question.

Mike

[CBedo]

I agree as well, that a distribution based speed figure makes much more sense, but unless you can integrate that into a larger probabilistic handicapping framework to use them, then they probably only end up confusing the issue.

[thaskalos]

Correct me if I'm wrong. I believe you are referring to the term, Triangulation.
If that is the case, you are correct about using various sources as well as different methods to estimate a speed rating.

The value of triangulation is to determine if convergence exists. Using the data sources (Bris, Equibase, and DRf figures), we would like to see evidence of the convergence of ratings. Sometimes convergence occurs, however, it is also likely it might not.

In fact, there are three possible outcomes that might result from a Triangulation strategy.

The first is the goal you are seeking convergence. That is agreement amongst the different sources or methods.

The second is inconsistency. That is the ratings may be inconsistent, not confirming but not contradictory.

The third is contradictory. The estimated ratings from differing methods and sources contradict each other.

In each case the estimated speed ratings have to be filtered through knowledge gleaned from the data, context, and understanding of the process.

I hope this answers your question.

MikeYes...this clarifies your argument completely. Thanks :ThmbUp:

[Greyfox]

Triangulate, or what ever you guys are calling it, horses are not racing cars.
The best predictors, are past performances. (That's what we can look at.)
Having said that, the best conditioned horse, who can overcome today's Pace, however he/she can do it, will and does bite that theory, and any others, in the balls.

[Charlie D]

the best conditioned horse, who can overcome today's Pace, however he/she can do it, will and does bite that theory, and any others, in the balls.




Sounds good to me.

[Fastracehorse]

Wasn't me backing Rachel at odds on Fast, wasn't me betting Vineyard Haven, as deafeat was out of the question . So i guess it must have been those people with the figs that make everything crystal clear.


..............but U don't want figs to pick u 1:9's : that should b self evident

fffastt

[Cratos]

Those most vocal against figs really have no clue how we used them.

If that is true, then the science of math and physics; and the practice of statistical applications have been wasted on a few people.

[Tom]

If that is true, then the science of math and physics; and the practice of statistical applications have been wasted on a few people.

How do those things tell you how I use figs?

[Cratos]

How do those things tell you how I use figs?

No, you tell me and the forum what the scientific basis of speed figures is if it is not math, physics, and statistics.

I am not demeaning or criticizing those who use speed figures, but for you to say “Those most vocal against figs really have no clue how we used them” is both perforated and disingenuous

[Cratos]

Aside from the methodology used to derive the figures, they are presented as a number, when it should be a distribution. All speed figures are subject to measurement errors, some more than others.

A speed figure is comprised of
estimated speed figure = true speed figure + error

Only God knows the true speed figure, and the error has two components: random noise and systematic.

The random error affects the variability of the estimated speed rating. Whereas the systematic error affects the mean vaule of the estimated speed rating. This results in a bias.

One of the best things you can do to deal with measurement errors, especially systematic errors, is to use multiple measures of the same estimated speed figure. Especially if the different measures don't share the same systematic errors, you will be able to triangulate across the multiple estimates and get a more accurate sense of what's going on.

Mike

I don't disagree with you, but what distribution would you suggest?

[Tom]

No, you tell me and the forum what the scientific basis of speed figures is if it is not math, physics, and statistics.

I am not demeaning or criticizing those who use speed figures, but for you to say “Those most vocal against figs really have no clue how we used them” is both perforated and disingenuous

I can't tell you - if you read what I wrote, I said THEY have no clue how we use them. Do you know how I use them? No, you don't. Neither do those who have posted nonsense here about following the last race fig. Do you know how CJ uses them? How Andicap uses them? No, you don't. I can tell you that statistics and physics don't enter into using the figs for me. And physics dosen't enter into making them for me, either. Never.

[TrifectaMike]

I don't disagree with you, but what distribution would you suggest?

I suggest a normal distribution. However, I would also suggest that it be used in a regional, and not in a global sense. By this I mean that the distribution is normal, but the variance is a function of the magnitude of the speed rating.

For example:

Using a Beyer type scale for the speed rating, we would have a normal distribution with a regional sigma.

Beyers between 91 - 100 sigma1
Beyers between 81 - 90 sigma2

and so on.

And for practical reasons you can truncate the distributions.

Mike

[Robert Goren]

The problem with speed figures is the daily variant. Some times deviance from from whatever base you use is pretty constant, but other time it all over the place. If you make the numbers you know that, if you get them from somebody else you don't. Some days you see things like +5, +4, +4, +6, +5, etc. Other days you see things like +5, -4, +4, -10, +11, -3, +8 etc. In the first case you can be pretty sure it is someplace around +5, in the second case you can't sure of anything. The second case happens a lot. Sometimes pace explain away some of it, but there are days when try as you may there is no answer. That being said, making your own numbers is a lot of work. Really a lot of work if you do it right and deal with more than one circuit.

[illinoisbred]

The problem with speed figures is the daily variant. Some times deviance from from whatever base you use is pretty constant, but other time it all over the place. If you make the numbers you know that, if you get them from somebody else you don't. Some days you see things like +5, +4, +4, +6, +5, etc. Other days you see things like +5, -4, +4, -10, +11, -3, +8 etc. In the first case you can be pretty sure it is someplace around +5, in the second case you can't sure of anything. The second case happens a lot. Sometimes pace explain away some of it, but there are days when try as you may there is no answer. That being said, making your own numbers is a lot of work. Really a lot of work if you do it right and deal with more than one circuit.
I agree RG. Sometimes pace does explain away some. So can some 2 or 3 yr old races where the winner explodes to a huge new top. Conversely, some slower races may just involve slow-subpar performers. Many fields of 5,000 to 7,500 older f+m claimers here are seriously subpar.

[PhantomOnTour]

I agree RG. Sometimes pace does explain away some. So can some 2 or 3 yr old races where the winner explodes to a huge new top. Conversely, some slower races may just involve slow-subpar performers. Many fields of 5,000 to 7,500 older f+m claimers here are seriously subpar.
Speaking of 2yr olds.....do you have/use 2yr old pars? I never use them.

[illinoisbred]

Speaking of 2yr olds.....do you have/use 2yr old pars? I never use them.
At Arlington I do have them for Maiden Specials-open and state-bred at 5 fur. only. I never use those races for variant sake-just like to see how the field measured up against the par.

I have pars for more classes at Hawthorne and will occasionally use some late,late in the year. Never use them for 2 yr olds routing.

[PhantomOnTour]

At Arlington I do have them for Maiden Specials-open and state-bred at 5 fur. only. I never use those races for variant sake-just like to see how the field measured up against the par.
I have pars for more classes at Hawthorne and will occasionally use some late,late in the year. Never use them for 2 yr olds routing.
Right on :ThmbUp:

I find most 2yr old sprints (this is strictly in NY) go fast early and slow late. In other words, their pace 'par' exceeds the Clm10 par for older horses but falls way short on final time. This is for 5.5f only.
I use the 'pars' like you. Just as a comparison to or enforcer of my TV for the day.

[illinoisbred]

Right on :ThmbUp:

I find most 2yr old sprints (this is strictly in NY) go fast early and slow late. In other words, their pace 'par' exceeds the Clm10 par for older horses but falls way short on final time. This is for 5.5f only.
I use the 'pars' like you. Just as a comparison to or enforcer of my TV for the day.
I've noticed the same here.In fact,2 yr old fillies in past years exceeding not only their male 2yr old counterparts but older males as well on their pace par.

[Fastracehorse]

No, you tell me and the forum what the scientific basis of speed figures is if it is not math, physics, and statistics.

I am not demeaning or criticizing those who use speed figures, but for you to say “Those most vocal against figs really have no clue how we used them” is both perforated and disingenuous

............................from our discussions? J/K.

You criticize speed figs for a 10 % +/- error.

I incorporate them because of this error %.

Watch a horse race: they are not aerodynamically engineered Blue Flames trying to break the speed of sound on a salt flat.

Horses are held, they are 'asked', they compete, they shy, they have problems, they rarely run a straight line, they drift, they get bumped, they don't like the footing.

fffastt

[Fastracehorse]

The problem with speed figures is the daily variant. Some times deviance from from whatever base you use is pretty constant, but other time it all over the place. If you make the numbers you know that, if you get them from somebody else you don't. Some days you see things like +5, +4, +4, +6, +5, etc. Other days you see things like +5, -4, +4, -10, +11, -3, +8 etc. In the first case you can be pretty sure it is someplace around +5, in the second case you can't sure of anything. The second case happens a lot. Sometimes pace explain away some of it, but there are days when try as you may there is no answer. That being said, making your own numbers is a lot of work. Really a lot of work if you do it right and deal with more than one circuit.

......making #'s ain't that much work and the variant really isn't a problem.

fffastt

[cj]

......making #'s ain't that much work and the variant really isn't a problem.

fffastt

If you believe that, the numbers you are making "ain't much" either.

[gm10]

I suggest a normal distribution. However, I would also suggest that it be used in a regional, and not in a global sense. By this I mean that the distribution is normal, but the variance is a function of the magnitude of the speed rating.

For example:

Using a Beyer type scale for the speed rating, we would have a normal distribution with a regional sigma.

Beyers between 91 - 100 sigma1
Beyers between 81 - 90 sigma2

and so on.

And for practical reasons you can truncate the distributions.

Mike

Looking at the horse population overall, this works quite well in the majority of cases, but it doesn't work so well in the tails of the distribution (= the very good and the very bad horses).

The exponential distribution is a slightly better option, but still far from perfect. If you use a member of the Gamma or Weibull families, you will be able to find a reasonable estimate for the speed figure and running time distributions of most horses. If nothing works, go non-parametric.

Fitting this into a larger probabilistic framework can be done through various ways. You can calculate the probability of each horse running faster than par time/figure, for example P(X>=80) where X is the speed figure RV and 80 is class par. You can work with order statistics (although this becomes more difficult if you can't assume a normal distribution). You can use them as input variables for broader models, such as a multinomial logit model.

The making of speed figures can be scientific but I suspect that most makers do at least some minor fiddling to make them look more reasonable. The way you use speed figures, however, is up to you and can be as scientific as you like.

[illinoisbred]

......making #'s ain't that much work and the variant really isn't a problem.

fffastt
Many days aren't difficult,but there are days that are tough. The last 2 days at Arlington have been quite difficult for both the turf and poly. Whether one uses strictly par times,or a projection method (I use a combination of both),these 2 days were amongst the toughest in the last several years.

[cj]

Many days aren't difficult,but there are days that are tough. The last 2 days at Arlington have been quite difficult for both the turf and poly. Whether one uses strictly par times,or a projection method (I use a combination of both),these 2 days were amongst the toughest in the last several years.

Arlington is always among the tracks with the most tough days. I don't really know what they are doing there.

[illinoisbred]

Arlington is always among the tracks with the most tough days. I don't really know what they are doing there.
The poly has been really difficult this year,and now with an abundance of turf races there's fewer poly races to make a variant determination. Throw in a couple of ungodly slow- paced races and the job does become difficult.This year's expansion of turf racing has helped when it comes to turf variant. Could do without the cheap nw2/3 on the turf and fewer of the endless starter allowances going at 5 furlongs. I guess they card what they can fill-they're clearly having a rough time filling many of the polytrack races.

[TrifectaMike]

Looking at the horse population overall, this works quite well in the majority of cases, but it doesn't work so well in the tails of the distribution (= the very good and the very bad horses).

The exponential distribution is a slightly better option, but still far from perfect. If you use a member of the Gamma or Weibull families, you will be able to find a reasonable estimate for the speed figure and running time distributions of most horses. If nothing works, go non-parametric.

Fitting this into a larger probabilistic framework can be done through various ways. You can calculate the probability of each horse running faster than par time/figure, for example P(X>=80) where X is the speed figure RV and 80 is class par. You can work with order statistics (although this becomes more difficult if you can't assume a normal distribution). You can use them as input variables for broader models, such as a multinomial logit model.



The making of speed figures can be scientific but I suspect that most makers do at least some minor fiddling to make them look more reasonable. The way you use speed figures, however, is up to you and can be as scientific as you like.

I agree the tails can be a source of error. However, it'an acceptable error. I also agree that the single one-parameter expontential distribution can be shown this data quite nicely. I also agree that a Gamma Distribution on a theoretical basis would be a better choice, but also more difficult to work with, especially if you are using Order Statistics as I do.

For those that are unfamiliar with Order Statistics. let me define it for you in non-mathemetaical terms.

Order Statistics is that body of knowledge which utilizes the rank or order of an observation as well as its magnitude. It combines the techniques of conventional statistics (which consider the magnitude of the observations) with those of rank order statistics ( which consider only the relative rank whether or not the original observations were measured on a ordinal scale).

I know many may be asking what does this have to do with picking winners, and I can understand that concern. With so many using Regression, NN's and GA's, why would anyone go in a different direction.

I'd like to commend you in describing the examples of a probabilistic framework. You've packed lots of good information in a single paragraph.

Mike

[Cratos]

I suggest a normal distribution. However, I would also suggest that it be used in a regional, and not in a global sense. By this I mean that the distribution is normal, but the variance is a function of the magnitude of the speed rating.

For example:

Using a Beyer type scale for the speed rating, we would have a normal distribution with a regional sigma.

Beyers between 91 - 100 sigma1
Beyers between 81 - 90 sigma2

and so on.

And for practical reasons you can truncate the distributions.

Mike

Okay you use a normal distribution and access the noise around the reference point at being 1 sigma. Correct me if I am wrong, but at the Beyer scale of 91-100 aren’t you looking at about 68% of the distribution data; what about the tails?

[Cratos]

............................from our discussions? J/K.

You criticize speed figs for a 10 % +/- error.

I incorporate them because of this error %.

Watch a horse race: they are not aerodynamically engineered Blue Flames trying to break the speed of sound on a salt flat.

Horses are held, they are 'asked', they compete, they shy, they have problems, they rarely run a straight line, they drift, they get bumped, they don't like the footing.

fffastt

Fast, your retort is out there in the “wild blue yonder” aerodynamically speaking. You don’t need to know astrophysics to calculate the movement of a race horse or any object in motion, but it helps that you understand the basic principles of physics.

The speed of a racehorse is the magnitude of its velocity and that is physics. Calculating averages and variants are statistics. And the tool to the work is math.

You might not like the disciplines of physics, statistics, and math; and you might have a methodology that excludes their use in your handicapping and to that I say bravo you have circumvented the necessary.

[Cratos]

Looking at the horse population overall, this works quite well in the majority of cases, but it doesn't work so well in the tails of the distribution (= the very good and the very bad horses).

The exponential distribution is a slightly better option, but still far from perfect. If you use a member of the Gamma or Weibull families, you will be able to find a reasonable estimate for the speed figure and running time distributions of most horses. If nothing works, go non-parametric.

Fitting this into a larger probabilistic framework can be done through various ways. You can calculate the probability of each horse running faster than par time/figure, for example P(X>=80) where X is the speed figure RV and 80 is class par. You can work with order statistics (although this becomes more difficult if you can't assume a normal distribution). You can use them as input variables for broader models, such as a multinomial logit model.

The making of speed figures can be scientific but I suspect that most makers do at least some minor fiddling to make them look more reasonable. The way you use speed figures, however, is up to you and can be as scientific as you like.

I believe that you are on to something with the exponential distribution and I would think that you are using the Poisson process.

However if the racetracks would weigh and publish the pre-race body weight of the horse then calculating speed would be much easier because now you would have the “mass” variable identified.

To me this is a major problem and many figures makers use a contrived track surface variation to explain away the plus/minus variation in a horses speed output; when it could be just the horse’s energy change.

[Fastracehorse]

If you believe that, the numbers you are making "ain't much" either.

I do a mental calculation for every horse that runs on every card I play.

I've been doing it for years now, and requires concentration to handicap, but it's a labour of love.

But you already knew that.

fffastt

[Fastracehorse]

Many days aren't difficult,but there are days that are tough. The last 2 days at Arlington have been quite difficult for both the turf and poly. Whether one uses strictly par times,or a projection method (I use a combination of both),these 2 days were amongst the toughest in the last several years.

...............my point is that I don't sweat it; I'm an advocate of good enoughness; I'm comfortable with a +/- 10 % error in my figs.

That's all anybody needs to gain a potent advantage.

Besides, most people know the game just isn't about figs.

fffastt

[Fastracehorse]

Fast, your retort is out there in the “wild blue yonder” aerodynamically speaking. You don’t need to know astrophysics to calculate the movement of a race horse or any object in motion, but it helps that you understand the basic principles of physics.

The speed of a racehorse is the magnitude of its velocity and that is physics. Calculating averages and variants are statistics. And the tool to the work is math.

You might not like the disciplines of physics, statistics, and math; and you might have a methodology that excludes their use in your handicapping and to that I say bravo you have circumvented the necessary.

..........in Canada we measure the speed of a racehorse using the metric system: METERS PER SECOND SQUARED :)

fffastt

[TrifectaMike]

..........in Canada we measure the speed of a racehorse using the metric system: METERS PER SECOND SQUARED :)

fffastt

In Canada you also have different physical laws. Per Second Squared? In which universe?

Mike

[TrifectaMike]

Okay you use a normal distribution and access the noise around the reference point at being 1 sigma. Correct me if I am wrong, but at the Beyer scale of 91-100 aren’t you looking at about 68% of the distribution data; what about the tails?

I'm not certain I understand you question. Perhaps my description wasn't clear enough. I'll take another stab at it.

When I refer to Beyers of 90-100 with a regional sigma (90-100 is for example only), I am saying that you'll have a family of normal distributions with an associated dispersion.

And when I refer another range of Beyers. I define another family of normal distributions with associated dispersion.

And if I were using the above distributions, I would truncate the distribution to bound both from below and above.

May be this clears it a bit.

Mike

[Greyfox]

..........in Canada we measure the speed of a racehorse using the metric system: METERS PER SECOND SQUARED :)

fffastt

Almost correct. Canada has always been ahead of the curve in some ways.
It's meters per second squared X IWWBT/ K.

(IWWBT = idiots who would believe that. K is a constant.)

[TrifectaMike]

I believe that you are on to something with the exponential distribution and I would think that you are using the Poisson process.

I don't believe that is what he saying. A Poisson process is a time dependent process and has a Poisson distribution.

The only connection I am aware of betwween a Poisson process and the exponential distribution is that: The waiting time probability distribution is an exponential distribution.

I don't believe this relates to the discussion at hand.

Mike

[Tom]

How deep are you going to go on a day with two route races?

[classhandicapper]

How deep are you going to go on a day with two route races?

At a certain point I think you have to admit to yourself that between track maintenance between races, changes in wind direction and intensity from race to race, some sections of the track being in sunlight and others not, varying run ups, extreme complications related to pace and between call action, bumping at the start that could slow all the horse down and an endless list of other issues that impact time, the upside of trying to be more accurate eventually reaches a point diminishing returns that makes it silly to pursue when there are other areas that will yield way more value.

As long as you have reasonably accurate figures and don't make any bets based on gross errors or misunderstandings you are in safe and productive water.

[Fastracehorse]

In Canada you also have different physical laws. Per Second Squared? In which universe?

Mike


........from 6th to 1st; isn't that m/(sec) squared - at least from the 2 points of the move?


fffastt

[Fastracehorse]

Almost correct. Canada has always been ahead of the curve in some ways.
It's meters per second squared X IWWBT/ K.

(IWWBT = idiots who would believe that. K is a constant.)

............the teeth of the leg trap comes down on the fox at a force = to mass x asseleration.

fffastt

[Fastracehorse]

At a certain point I think you have to admit to yourself that between track maintenance between races, changes in wind direction and intensity from race to race, some sections of the track being in sunlight and others not, varying run ups, extreme complications related to pace and between call action, bumping at the start that could slow all the horse down and an endless list of other issues that impact time, the upside of trying to be more accurate eventually reaches a point diminishing returns that makes it silly to pursue when there are other areas that will yield way more value.

As long as you have reasonably accurate figures and don't make any bets based on gross errors or misunderstandings you are in safe and productive water.

........there aren't many avenues to persue that are more profitable than figs; despite the factors that lead the path to perfection in creating speed figs unnavigatable - good enoughness.

Even statistics have a probability +/- error rate.

fffastt

[thaskalos]

........there aren't many avenues to persue that are more profitable than figs. We don't need MANY avenues...we just need one.

[Charlie D]

We don't need MANY avenues...we just need one.

K.I.S.S would be the one avenue i would advise.

[classhandicapper]

........there aren't many avenues to persue that are more profitable than figs; despite the factors that lead the path to perfection in creating speed figs unnavigatable - good enoughness.

Even statistics have a probability +/- error rate.

fffastt

I find profits without figures. To be quite honest, the most profitable period of my life came when I used no figures at all.

I use figures now (both pace and final time) because I want to make sure the horse I'm betting is fast enough to win if it fires its "A" race and to help me interpret the results of races. With some exceptions, I find that most horses with a solid edge on figures are usually heavily bet.

Most of my profits come from exploiting biases (both on the day of the race and when the horses come back), knowing that a horse is very likely to improve or fire a big race based on a trainer pattern that is not well known yet, finding a horse that is very likely to improve it figures because it has been facing superior horses (hidden form and class drops) and other things that also require work and research.

If I was spending all my free time trying to determine if a particular horse really ran a 94 or a 97 last time out, I think I'd probably miss out on a lot of other important things. As long as I know the horse has been running in the mid 90s and I trust the accuracy of the figure maker, I think that's good enough.

(I used to make my own figures but I trust that CJs figures are as good or better than the ones I made for myself. The style of figures is identical and he covers way more tracks than I would have time for.)

[illinoisbred]

I find profits without figures. To be quite honest, the most profitable period of my life came when I used no figures at all.

I use figures now (both pace and final time) because I want to make sure the horse I'm betting is fast enough to win if it fires its "A" race and to help me interpret the results of races. With some exceptions, I find that most horses with a solid edge on figures are usually heavily bet.

Most of my profits come from exploiting biases (both on the day of the race and when the horses come back), knowing that a horse is very likely to improve or fire a big race based on a trainer pattern that is not well known yet, finding a horse that is very likely to improve it figures because it has been facing superior horses (hidden form and class drops) and other things that also require work and research.

If I was spending all my free time trying to determine if a particular horse really ran a 94 or a 97 last time out, I think I'd probably miss out on a lot of other important things. As long as I know the horse has been running in the mid 90s and I trust the accuracy of the figure maker, I think that's good enough.

(I used to make my own figures but I trust that CJs figures are as good or better than the ones I made for myself. The style of figures is identical and he covers way more tracks than I would have time for.)
That's the big drawback with figures for me-having them for all the tracks horses ship in from. I make them for Arlington,Hawthorne,and Fair Grounds.If time wasn't an issue,I'd like to have them for Churchill, Keeneland,Oaklawn,Prairie Meadows, and Canterbury. Because of this, most of my serious action is confined to state-bred races.

[PhantomOnTour]

BRIS leads the way in this poll with OTHER coming in 2nd which is a surprise to me, as is the DRF/BEYER being a distant 3rd. Ya'll think this has anything to do with price? BRIS pp's are by and large free thru your ADW company or other links, and OTHER could be free if you are making your own figs. I wish that was a possible answer: I MAKE MY OWN. Ofcourse, RAGS are costly and they came in 4th.

Used to be that DRF/BEYER was the dominant one, but their pp's are costly compared to BRIS. BRIS leads the way but I read post after post on this board stating that their figs are among the worst....hmmm.

[illinoisbred]

Phantom-do you find may discrepancies between your own figures and the Bris figures? I certainly do here particularily with poly races and many,many turf races and routes in general.

[PhantomOnTour]

Phantom-do you find may discrepancies between your own figures and the Bris figures? I certainly do here particularily with poly races and many,many turf races and routes in general.
Me and BRIS differ a lot, but I don't do figs for any poly tracks and my turf figs aren't the traditional pace & speed numbers. I do daily figs for the NY circuit (not FL or AQU inner) and their prominent summer feeder tracks like Mth and CD and GP, if you can call them 'feeders'. I follow CD and GP because many ship to Bel and Sar.

For turf I basically do a pace-final fraction number (pace plus final fraction=final time but it's a lil different from that). I give a pace 'rating' that is Slow Avg or Fast and then a numerical final fraction rating (ex: a race will be rated F-102 which means fast pace with a 102 rated final fraction...par for me is always 100). Numerical pace ratings in turf routes routes just don't work for me as the splits vary wildly. I just wanna know if they went slow, avg or fast early to help me interpret the final fraction rating.

[Fingal]

Since the original question was about a LONE STAND ALONE number, I'll stick with Today's Racing Digest as that single vote. But any stand alone figure be it BRIS, Digest, home grown or whomever is only as good as one's ability to determine contenders & not even bother looking at those eliminated in that process.

Numbers help you quantify things-
Experience tells you what to quantify

[Fastracehorse]

We don't need MANY avenues...we just need one.

.....but it is my nature; I like to b included in every race - I play alot alot alot of horizontal exotics

fffastt

[Fastracehorse]

I find profits without figures. To be quite honest, the most profitable period of my life came when I used no figures at all.

I use figures now (both pace and final time) because I want to make sure the horse I'm betting is fast enough to win if it fires its "A" race and to help me interpret the results of races. With some exceptions, I find that most horses with a solid edge on figures are usually heavily bet.

Most of my profits come from exploiting biases (both on the day of the race and when the horses come back), knowing that a horse is very likely to improve or fire a big race based on a trainer pattern that is not well known yet, finding a horse that is very likely to improve it figures because it has been facing superior horses (hidden form and class drops) and other things that also require work and research.

If I was spending all my free time trying to determine if a particular horse really ran a 94 or a 97 last time out, I think I'd probably miss out on a lot of other important things. As long as I know the horse has been running in the mid 90s and I trust the accuracy of the figure maker, I think that's good enough.

(I used to make my own figures but I trust that CJs figures are as good or better than the ones I made for myself. The style of figures is identical and he covers way more tracks than I would have time for.)



.................you use figs because you've been BLOWN AWAY too many times without them.

2nd: Figs uncover longshots because they need to b adjusted.

3rd: They are 1 important aspect of many, sad to say many, but I am a comprehensive handicapper, ie, if I were a forward in hockey I'd have more than 1 move.

fffastt

[Fastracehorse]

Since the original question was about a LONE STAND ALONE number, I'll stick with Today's Racing Digest as that single vote. But any stand alone figure be it BRIS, Digest, home grown or whomever is only as good as one's ability to determine contenders & not even bother looking at those eliminated in that process.

Numbers help you quantify things-
Experience tells you what to quantify

.......is 2 quantify everything; i'm not allowed 2 gloss over the quantification part

fffastt

[classhandicapper]

.................you use figs because you've been BLOWN AWAY too many times without them.


fffastt

Definitely NOT!!!

I use them because in a small number of situations I have insufficient information to do a qualitative analysis of a field. In those instances, I know that figures will give me a better ballpark estimate of the quality than I could come up without them.

I went from a pure numbers guy 25-30 years ago (mostly successful with trainer information/patterns and not figures) to a class handicapper (very successful from all directions for about 10 years) to understanding the strengths and weaknesses of both approachs and trying to focus on the strengths of each when they seemed to be the most applicable.

So now I use pace and speed figures when it makes sense to use them but I basically disregard them when I feel certain I understand the abilities of the horses better than reflected in numbers.

[Cratos]

I'm not certain I understand you question. Perhaps my description wasn't clear enough. I'll take another stab at it.

When I refer to Beyers of 90-100 with a regional sigma (90-100 is for example only), I am saying that you'll have a family of normal distributions with an associated dispersion.

And when I refer another range of Beyers. I define another family of normal distributions with associated dispersion.

And if I were using the above distributions, I would truncate the distribution to bound both from below and above.

May be this clears it a bit.

Mike

I apologize and I misunderstood your original post; I now understand what you were stating.

[Cratos]

I don't believe that is what he saying. A Poisson process is a time dependent process and has a Poisson distribution.

The only connection I am aware of betwween a Poisson process and the exponential distribution is that: The waiting time probability distribution is an exponential distribution.

I don't believe this relates to the discussion at hand.

Mike

I believe you and GM10 have a stronger statistics background than me, but would not the number of strides a horse takes over a given race distance be the homogeneous Poisson process and would not it follow a Poisson distribution?

Therefore given that the Poisson distribution would be counting the number of discrete events (if my above assumption is correct) in a fixed time period and would this not relate or connect the Poisson distribution to the exponential distribution

What I am suggesting is that if X denotes the number of strides and Y is the time between strides; would not Y be exponentially distributed?

[gm10]

I don't believe that is what he saying. A Poisson process is a time dependent process and has a Poisson distribution.

The only connection I am aware of betwween a Poisson process and the exponential distribution is that: The waiting time probability distribution is an exponential distribution.

I don't believe this relates to the discussion at hand.

Mike

Not suggesting you do or don't, but you can model running times this way (horses arriving at the finish would be the Poisson process). I'm sure it works well, whether you want to invest the time & effort depends on what statistical plans you have with your figures!

I was thinking about your 'triangulating' of different speed figures this weekend. I can't make up my mind about it. The scale of 0-140 may be used by most figure makers, but this doesn't mean that their figures follow the same distribution of course. In the case of BSF I would think that they are not normally distributed. Ultimately, they are a linear function of running times if I'm not mistaken? Any insights?

[TrifectaMike]

I believe you and GM10 have a stronger statistics background than me, but would not the number of strides a horse takes over a given race distance be the homogeneous Poisson process and would not it follow a Poisson distribution?

Therefore given that the Poisson distribution would be counting the number of discrete events (if my above assumption is correct) in a fixed time period and would this not relate or connect the Poisson distribution to the exponential distribution

What I am suggesting is that if X denotes the number of strides and Y is the time between strides; would not Y be exponentially distributed?

You are absolutely correct. And I said as much in the post you replied to. However, my earlier post were not about how to compute ratings, but the measurement errors in the estimated ratings.

Mike

[TrifectaMike]

Not suggesting you do or don't, but you can model running times this way (horses arriving at the finish would be the Poisson process). I'm sure it works well, whether you want to invest the time & effort depends on what statistical plans you have with your figures!

I was thinking about your 'triangulating' of different speed figures this weekend. I can't make up my mind about it. The scale of 0-140 may be used by most figure makers, but this doesn't mean that their figures follow the same distribution of course. In the case of BSF I would think that they are not normally distributed. Ultimately, they are a linear function of running times if I'm not mistaken? Any insights?

True, the different figures might have different distributions assocated with figures. However, the dispersion about a particular figure will be normally distributed. Once again, I'm interested in measurement errors.

Mike

[Fastracehorse]

Definitely NOT!!!

I use them because in a small number of situations I have insufficient information to do a qualitative analysis of a field. In those instances, I know that figures will give me a better ballpark estimate of the quality than I could come up without them.

I went from a pure numbers guy 25-30 years ago (mostly successful with trainer information/patterns and not figures) to a class handicapper (very successful from all directions for about 10 years) to understanding the strengths and weaknesses of both approachs and trying to focus on the strengths of each when they seemed to be the most applicable.

So now I use pace and speed figures when it makes sense to use them but I basically disregard them when I feel certain I understand the abilities of the horses better than reflected in numbers.


............and class definitely is; but with less clarity than speed figs.

fffastt

[classhandicapper]

............and class definitely is; but with less clarity than speed figs.

fffastt

That's where we differ.

IMO sometimes CLASS is clearer, but in others figures are.

[Cratos]

You are absolutely correct. And I said as much in the post you replied to. However, my earlier post were not about how to compute ratings, but the measurement errors in the estimated ratings.

Mike

Mike, I understand your objective here in this thread with your posts, but without going into how the speed rating or the speed figure is computed, the speed rating or the speed figure becomes what Macbeth described as a tale “full of sound and fury, signifying nothing.”

[TrifectaMike]

Mike, I understand your objective here in this thread with your posts, but without going into how the speed rating or the speed figure is computed, the speed rating or the speed figure becomes what Macbeth described as a tale “full of sound and fury, signifying nothing.”

Point taken.

Let me relate a story to you. Back in a previous life as a researcher, I and a small group were tasked with developing techniques for on-board real time calibration of the Space Shuttle Navigation System. We developed a slew of Kalman filters for our specific purpose (cutting edge at the time). We also defined a few Lyapunov functions for the design of robust and adaptive controllers (We also used Lyaponov functions and Lyaponov Stability for system identification and specification for certain parameters). Much of the work still remains classified and is used in many missile systems today.

Now, as much as love and respect engineers, they were at times an odd lot. I had accelerometer, gyro and gimbal designers constantly asking if I needed to know how their instruments were designed. I'd normally respond...not really. I know how they work. I'll simply make observations. Of course I'd get a WTF look. That would last until our software designers would receive their algorithms that made their instruments perform much better.

My point is that, I did not need to know how the observations were generated. What I had to determine is how I could efficiently and accurately identify the biases, first order and higher terms. My interest was not to build a better instrument, my interest was to make a state of the art instrument better.

Relate this to the estimated speed ratings, and you'll get a better appreciation of what I attempt to accomplish, and what I speak to in this thread.

Please, don't take this personally. In fact, I find your posts about using engineering and physical laws quite interesting. And for the most part accurate.

Mike

[Cratos]

Point taken.

Let me relate a story to you. Back in a previous life as a researcher, I and a small group were tasked with developing techniques for on-board real time calibration of the Space Shuttle Navigation System. We developed a slew of Kalman filters for our specific purpose (cutting edge at the time). We also defined a few Lyapunov functions for the design of robust and adaptive controllers (We also used Lyaponov functions and Lyaponov Stability for system identification and specification for certain parameters). Much of the work still remains classified and is used in many missile systems today.

Now, as much as love and respect engineers, they were at times an odd lot. I had accelerometer, gyro and gimbal designers constantly asking if I needed to know how their instruments were designed. I'd normally respond...not really. I know how they work. I'll simply make observations. Of course I'd get a WTF look. That would last until our software designers would receive their algorithms that made their instruments perform much better.

My point is that, I did not need to know how the observations were generated. What I had to determine is how I could efficiently and accurately identify the biases, first order and higher terms. My interest was not to build a better instrument, my interest was to make a state of the art instrument better.

Relate this to the estimated speed ratings, and you'll get a better appreciation of what I attempt to accomplish, and what I speak to in this thread.

Please, don't take this personally. In fact, I find your posts about using engineering and physical laws quite interesting. And for the most part accurate.

Mike

It was just the engineer in me speaking aloud

[bobphilo]

CJ's figures, for both speed and pace.Bob

[gm10]

True, the different figures might have different distributions assocated with figures. However, the dispersion about a particular figure will be normally distributed. Once again, I'm interested in measurement errors.

Mike

How do you define the measurement error? Do you have a number that you accept as the 'true speed figure', against which you measure errors?

[TrifectaMike]

How do you define the measurement error? Do you have a number that you accept as the 'true speed figure', against which you measure errors?

Normally, I don't disclose the techniques I utilize, but in this case I'll make an exception. I use the following to determine the "true speed figure" an Externally Bayesian Pooling rule. That is all I will disclose. Have fun, and compute away.

Mike

[Fastracehorse]

.......of speed figs.

A 3yo colt breaks his Maiden with a Beyer of 78 and is now running vs. winners today. He won going away after rating midpack and making a big move. He had more in the tank.

The fave in this race ran a 90 Beyer on the lead but just failed to hold - his 2nd effort in a row of this type.

Now suppose U had a more accurate determinator of ability than the Beyer; and you have the first colt running a 76.2 - and the 2nd a 91.3 - a slighltly greater separation.

Is the real important point the discrepancy in the measuring sticks of the two horses or how these horses are going to run today?

I say the latter; we have an estimate of ability - the tote likes the speed horse - but of course there are many factors at play here beyond the 2 measurements of past ability; because this obviously isn't just a 2 horse race.

1st off; knowing the nature of tb's, ie, being a form handicapper, the discrepancy in figs may not be relevant: horses improve - horses regress.

But if U believe in them steadfastly ( figs ), U may agree with the tote.

This is not a static game, very often, it is organic, meaning to me that there are a host of factors we cannot measure finitely; like: by how much will the favorite regress?

But with the two set of figures we can make a guess to what might happen; we have the good enough figs with a larger error rate or we have the highly accurate figs that try hard to measure an organic game with variables that are impossible to quantify in this race, or in most races.

Yes, the highly accurate speed figs tell U that the fave would probably win today if he ran exactly as he did last race; but the Beyer does too - is the discrepancy relevant?; especially since the error rate between the 2 figs are essentially neglible; and even more so on the stage of an organic game called thoroughbred horseracing.

fffastt

[thaskalos]

Even the most well-crafted pace and speed figures have in them a greater margin of error than many players realize. This margin of error is not an indictment against the figure maker, and his level of skill. It has more to do with the nature of the game.

This game has not yet advanced to the level where precise measurement of the performance of the horses is possible.

At this point and time...accurate speed figure interpretation is much more important than the slight discrepancies in the figures themselves, IMO.

[gm10]

Normally, I don't disclose the techniques I utilize, but in this case I'll make an exception. I use the following to determine the "true speed figure" an Externally Bayesian Pooling rule. That is all I will disclose. Have fun, and compute away.

Mike

ok thanks

[PhantomOnTour]

Even the most well-crafted pace and speed figures have in them a greater margin of error than many players realize. This margin of error is not an indictment against the figure maker, and his level of skill. It has more to do with the nature of the game.

This game has not yet advanced to the level where precise measurement of the performance of the horses is possible.

At this point and time...accurate speed figure interpretation is much more important than the slight discrepancies in the figures themselves, IMO.
In other parts of the world it has, but not here in America. Outside of Kee does any N American track use trackus or something similar to give the exact number of feet a horse traveled?
You may be speaking of something else entirely when you say 'precise measurement isn't possible'...if so I apologize. When I am able to get the exact amount of distance a horse traveled and accurate beaten lengths at the fractional splits I think I'll have all I need to truly measure a performance.

[Tom]

In other parts of the world it has, but not here in America. Outside of Kee does any N American track use trackus or something similar to give the exact number of feet a horse traveled?


What a joke Trakus has turned out to be. So what is KEE and WO use it - no one else does, so it's usefulness is pretty much limited. And where it is uses, do they really give you the horses time or do they still use a beaten length constant? The distance stuff is good, but I don't think we are getting the real times.

[thaskalos]

In other parts of the world it has, but not here in America. Outside of Kee does any N American track use trackus or something similar to give the exact number of feet a horse traveled?
You may be speaking of something else entirely when you say 'precise measurement isn't possible'...if so I apologize. When I am able to get the exact amount of distance a horse traveled and accurate beaten lengths at the fractional splits I think I'll have all I need to truly measure a performance.In Andy Beyer's book "Beyer On Speed" - which was published in 1993 - it was revealed that the technology existed which would finally bring handicapping into the 21st century. Micro chips would be placed in the bridles of the horses, which would trigger sensors that would be placed at the different points of call...finally giving the horseplayers the accurate individual horse velocity ratings that we so badly need...in order to properly deal with the complexities of this game.

That was 17 years ago...what seems to be the hold up?

Look at the tools that the horseplayer has to work with now.

I was recently handicapping a $10,000 sprint at Arlington Park...and spotted a horse whose last race had the following fractional times:

23.2 45.4 111.0

And the running line of this horse indicated that the horse had somehow registered a 7.5 length gain during this 22.2 second fraction.

These are the running lines that we base our intricate pace figures on...

[TrifectaMike]

.......of speed figs.

A 3yo colt breaks his Maiden with a Beyer of 78 and is now running vs. winners today. He won going away after rating midpack and making a big move. He had more in the tank.

The fave in this race ran a 90 Beyer on the lead but just failed to hold - his 2nd effort in a row of this type.

Now suppose U had a more accurate determinator of ability than the Beyer; and you have the first colt running a 76.2 - and the 2nd a 91.3 - a slighltly greater separation.

Is the real important point the discrepancy in the measuring sticks of the two horses or how these horses are going to run today?

I say the latter; we have an estimate of ability - the tote likes the speed horse - but of course there are many factors at play here beyond the 2 measurements of past ability; because this obviously isn't just a 2 horse race.

1st off; knowing the nature of tb's, ie, being a form handicapper, the discrepancy in figs may not be relevant: horses improve - horses regress.

But if U believe in them steadfastly ( figs ), U may agree with the tote.

This is not a static game, very often, it is organic, meaning to me that there are a host of factors we cannot measure finitely; like: by how much will the favorite regress?

But with the two set of figures we can make a guess to what might happen; we have the good enough figs with a larger error rate or we have the highly accurate figs that try hard to measure an organic game with variables that are impossible to quantify in this race, or in most races.

Yes, the highly accurate speed figs tell U that the fave would probably win today if he ran exactly as he did last race; but the Beyer does too - is the discrepancy relevant?; especially since the error rate between the 2 figs are essentially neglible; and even more so on the stage of an organic game called thoroughbred horseracing.

fffastt

If you believe that, I have proposition to make.

I'll produce an oddsline using Bri Speed Figures and constrast that with CJ's (FYI, I have never used CJ's figures), and let's see if there exist a difference on a per race basis. I'll use no other factor than the figures.

In fact, why not contrast the oddsline against any other... they can use as many factors as they wish.

All I would need to do this is speed figures from CJ. He can pick the track and date.

Mike

[Fastracehorse]

If you believe that, I have proposition to make.

I'll produce an oddsline using Bri Speed Figures and constrast that with CJ's (FYI, I have never used CJ's figures), and let's see if there exist a difference on a per race basis. I'll use no other factor than the figures.

In fact, why not contrast the oddsline against any other... they can use as many factors as they wish.

All I would need to do this is speed figures from CJ. He can pick the track and date.

Mike

...........................and I hate odds lines; they aren't relevant to me.

How bout U and me do selections at Belmont on Sundays for the rest of the meet.

I don't want it to be too time consuming.

fffastt

[TrifectaMike]

...........................and I hate odds lines; they aren't relevant to me.

How bout U and me do selections at Belmont on Sundays for the rest of the meet.

I don't want it to be too time consuming.

fffastt

Don't do selections...that's a foreign concept to me.

That said. I think you misunderstood my proposition. I wanted to use CJ's figures for a day to contrast against the Bris figures.

Mike

[bisket]

If you believe that, I have proposition to make.

I'll produce an oddsline using Bri Speed Figures and constrast that with CJ's (FYI, I have never used CJ's figures), and let's see if there exist a difference on a per race basis. I'll use no other factor than the figures.

In fact, why not contrast the oddsline against any other... they can use as many factors as they wish.

All I would need to do this is speed figures from CJ. He can pick the track and date.

Mike
i'm a trifecta player, and the first thought that came to mind is: what are the horses odds? i see a spot for a profitable exotic with the horse with the mid 70's fig. is there a good chance for a fast pace? whats the works look like and what were his (horse with mid 70's fig) races leading into this one. thats the angle i look for: what's the horse going to do for my tri. if odds are high enough he helps if he finishes first, second, or third....

[Greyfox]

Don't do selections...that's a foreign concept to me.

That said. I think you misunderstood my proposition. I wanted to use CJ's figures for a day to contrast against the Bris figures.

Mike

Seems to me to be unfair to both services.
1. sample too small
2. samplers too few
3. inter-rater reliability?

If you do something like that between the two of you privately, there is nothing wrong. But letting the general public on this board know which service you are using is not fair under such limitations.

[TrifectaMike]

...........................and I hate odds lines; they aren't relevant to me.

How bout U and me do selections at Belmont on Sundays for the rest of the meet.

I don't want it to be too time consuming.

fffastt

I envy guys like you. You obviously have abilities beyond mine. You can be objective and make selections. Unfortunately I'm a subjective kind of guy and I'm totally lost without my probabilities.

Well, I guess that's life. Some like you have "it", and others like me don't.

Good luck
Mike

[TrifectaMike]

Seems to me to be unfair to both services.
1. sample too small
2. samplers too few
3. inter-rater reliability?

If you do something like that between the two of you privately, there is nothing wrong. But letting the general public on this board know which service you are using is not fair under such limitations.

I agree, but you are making too much of my proposition. I just wanted to demonstrate how difference in figs can be amplified., and not as a competition.

Mike

[Fastracehorse]

Don't do selections...that's a foreign concept to me.

That said. I think you misunderstood my proposition. I wanted to use CJ's figures for a day to contrast against the Bris figures.

Mike

.....and your wrong.

J/K,

how do pace bets if U don't make selections? ( accidental pun )

I use speed figs too, not CJ's, but I adjust Beyers.

Let's do the Sunday Belmont thing - and when I beat you I promise U I'll teach U something :) And I will beat U, probably pretty damn bad


fffastt

[TrifectaMike]

Actually I'll scratch the idea. I don't know what I was thinking.

Mike

[mistergee]

.....and your wrong.

J/K,

how do pace bets if U don't make selections? ( accidental pun )

I use speed figs too, not CJ's, but I adjust Beyers.

Let's do the Sunday Belmont thing - and when I beat you I promise U I'll teach U something :) And I will beat U, probably pretty damn bad


fffastt
have you ever tried using the beyer pace #s

[Fastracehorse]

...................that way U are guaranteed 2nd :)

fffastt

[Fastracehorse]

have you ever tried using the beyer pace #s

.....but I have a theory on the subject.

fffastt

[Greyfox]

...................that way U are guaranteed 2nd :)

fffastt

:lol: :lol: Didn't Clint Eastwood say "Make my day."

(More seriously, I use my own figs. I load my own ammo.)

[TrifectaMike]

.....and your wrong.

J/K,

how do pace bets if U don't make selections? ( accidental pun )

I use speed figs too, not CJ's, but I adjust Beyers.

Let's do the Sunday Belmont thing - and when I beat you I promise U I'll teach U something :) And I will beat U, probably pretty damn bad


fffastt

Ok I concede. You can beat me, and it hurts so bad. So, now teach me something. I'll await your PM.

Thank you
Mike

[mistergee]

.....but I have a theory on the subject.

fffastt
the method to achieve them was described in one of his books but I dont remember which

[Fastracehorse]

:lol: :lol: Didn't Clint Eastwood say "Make my day."

(More seriously, I use my own figs. I load my own ammo.)

make my day punk

fffastt

[Fastracehorse]

Ok I concede. You can beat me, and it hurts so bad. So, now teach me something. I'll await your PM.

Thank you
Mike

......................it has to be a public lashing

it'll b good 4 u as far as furthering your game goes, and mine

fffastt

[Cratos]

How do you define the measurement error? Do you have a number that you accept as the 'true speed figure', against which you measure errors?

GM10, you have raised probably the most important question with the topic of error measurement within speed ratings and speed figures.

Therefore it should be clear to anyone who is producing a speed rating or a speed figure that the two most important aspects of a speed rating or a speed figure is the reliability and validity of their precision.

But what is reliability in this context? It is the ability to reproduce a measurement because as reliability degrades, the precision of the speed rating or the speed figure is reduced and the ability to monitor changes in the measurement of the speed rating or the speed figure becomes problematic.

This brings us to the validation of the a speed rating or a speed figure precision.. We do that through validity which refers to the agreement between the value of a speed rating or a speed figure measurement and their true value.

Quantification of validity comes by comparing the measurement of the speed rating or speed figure to the true value as close as possible. A degradation in validity will minimize the ability to differentiate relationships between a speed rating or a speed figure when studying them for future use.

[Fastracehorse]

....i have better luk w/ horses than women :)

ttyl


fffastt

[Cratos]

i'm a trifecta player, and the first thought that came to mind is: what are the horses odds? i see a spot for a profitable exotic with the horse with the mid 70's fig. is there a good chance for a fast pace? whats the works look like and what were his (horse with mid 70's fig) races leading into this one. thats the angle i look for: what's the horse going to do for my tri. if odds are high enough he helps if he finishes first, second, or third....

In betting racehorses under pari-mutuel conditions there are two types of odds. Tote board odds which are determined by the betting interactions from the bettors’ wagers and the odds set by the individual bettor to him/herself to make a wager.

Ideally the individual bettor’s odds should always be equal to or less than the tote board odds on the horse which the bettor desire to bet.

[TrifectaMike]

......................it has to be a public lashing

it'll b good 4 u as far as furthering your game goes, and mine

fffastt

Ahh. I'm disappointed. Finally, at least for a moment, I believed I could escape this probabilistic framework that I created for myself. Please help me further my "game"... at least give me a hint. Why not not be kind to an old guy like me. Please be my mentor. I know I can use one. I'll adjust Beyers. I'll do whatever you direct me to do. Any help from you, I'm certain will improve my game. In fact, I have so little game, it wouldn't take much to improve it.

I'll say it again, please help me.

Thank you
Mike

[TrifectaMike]

......................it has to be a public lashing

it'll b good 4 u as far as furthering your game goes, and mine

fffastt

Hurry up. I'm peeing my pants in anticpation. Tell me the SECRET!

Mike

[TrifectaMike]

......................it has to be a public lashing

it'll b good 4 u as far as furthering your game goes, and mine

fffastt

I figured as much. No help coming. I'll watch dancing with stars.

Disappointed,
Mike

[bisket]

In betting racehorses under pari-mutuel conditions there are two types of odds. Tote board odds which are determined by the betting interactions from the bettors’ wagers and the odds set by the individual bettor to him/herself to make a wager.

Ideally the individual bettor’s odds should always be equal to or less than the tote board odds on the horse which the bettor desire to bet.
this is probably what i place the most emphasis on in deciding which horse will be included in my wager, and whether i'll actually wager on the race or not. most times this analysis also involves whether or not i think the strongest entry in the race will run a good race. obviously in the example thats drf gave the biggest factor would be: do i think the race will set up for the horse with the 90 fig. most times if your playing a horse with long odds you are betting that the race will fall apart or it will unfold in an unlikely manner than the data says it will. most bettors don't like to bet against the data, but done correctly its not such a wild west style of wagering... calculating discipline!!!

[bisket]

beyer's figs have always been my choice of speed figures. for me drf has always been tops. i like to know what i'm getting in my handicapping, and can make my own adjustments when needed. i just don't like when fig makers take liberties with the data in attempts to make a number that is all encompassing. just give me a number thats compared adequately with the final times of the other races that day, and i'm happy. i just think the simple formula that beyer uses is still better than all these other adjustments others make with their figs. i'm capable of filling in the blanks after someone else takes care of the math for me. all i have to do is make sure i agree with the variant he uses every day.

[Fastracehorse]

Ahh. I'm disappointed. Finally, at least for a moment, I believed I could escape this probabilistic framework that I created for myself. Please help me further my "game"... at least give me a hint. Why not not be kind to an old guy like me. Please be my mentor. I know I can use one. I'll adjust Beyers. I'll do whatever you direct me to do. Any help from you, I'm certain will improve my game. In fact, I have so little game, it wouldn't take much to improve it.

I'll say it again, please help me.

Thank you
Mike

...................cuz I have no idea with the latter

fffastt

[Fastracehorse]

I figured as much. No help coming. I'll watch dancing with stars.

Disappointed,
Mike

never ever ever bet a longshot

they win at such low probabiity


J/K

I swear I read the above in a handicapping book once

fffastt

[Fastracehorse]

beyer's figs have always been my choice of speed figures. for me drf has always been tops. i like to know what i'm getting in my handicapping, and can make my own adjustments when needed. i just don't like when fig makers take liberties with the data in attempts to make a number that is all encompassing. just give me a number thats compared adequately with the final times of the other races that day, and i'm happy. i just think the simple formula that beyer uses is still better than all these other adjustments others make with their figs. i'm capable of filling in the blanks after someone else takes care of the math for me. all i have to do is make sure i agree with the variant he uses every day.

............that Beyers are good enough

ffffastt

[TrifectaMike]

There are moments in one's life, which are pivotal, and also life
changing. Last night was one of those moments.

I had a fortunate, for me, encounter with Fastracehorse@DRF. I can honestly
say the encounter has changed my view of the handicapping process.

I]d like to thank you publically for making such a difference in my
handicapping life and in such a short time. The impact you've had is
profound.

Why have I changed? It is because of statements like these:

Originally Posted by Fastracehorse@DRF
......making #'s ain't that much work and the variant really isn't a
problem.

fffastt

Originally Posted by Fastracehorse@DRF
..........in Canada we measure the speed of a racehorse using the metric system:
METERS PER SECOND SQUARED

fffastt

Originally Posted by Fastracehorse@DRF
2nd: Figs uncover longshots because they need to b adjusted.

fffastt


Originally Posted by Fastracehorse@DRF
...........................and I hate odds lines; they aren't relevant to me.
How bout U and me do selections at Belmont on Sundays for the rest of the meet.
I don't want it to be too time consuming.

fffastt

Originally Posted by Fastracehorse@DRF
how do pace bets if U don't make selections? ( accidental pun )
I use speed figs too, not CJ's, but I adjust Beyers.
Let's do the Sunday Belmont thing - and when I beat you I promise U I'll teach
U something And I will beat U, probably pretty damn bad

fffastt

Originally Posted by Fastracehorse@DRF
......................it has to be a public lashing
it'll b good 4 u as far as furthering your game goes, and mine

fffastt

Originally Posted by Fastracehorse@DRF
My best tip
never ever ever bet a longshot
they win at such low probabiity

J/K

I swear I read the above in a handicapping book once

fffastt

As I sit at my desk today, looking at the Belmont card, and
reflecting on the advice given by Fastracehorse@DRF, I have decided
not to wager today. Instead I am embarking on several studies.

I will study the Beyer Speed Rating.
I will study how speed is proportional to sec squared.
I will study the futility of oddslines.
I will study how to adjust Beyer Speed Ratings.
I will study how longshots win at win at such low probabiity.
I will study how to make Beyer type pace (I list this study last,
because I first have to complete the study on how speed is
proportional to sec squared. This may not seem as easy as it first
appears)
Most importantly, I will not make another wager until I complete
my studies. Not another wager until I rise to Fastracehorse@DRF's
skill level.

Once again, thank you, Fastracehorse@DRF, for pointing me in the
right direction. I know with your advice and direction there is a
profit in my future.

Mike

[thaskalos]

That's what this site is here for...

To lend us a helping hand when we are struggling.

With Fastracehorse's advise...you should be counting your profits in no time...

[Fastracehorse]

There are moments in one's life, which are pivotal, and also life
changing. Last night was one of those moments.

I had a fortunate, for me, encounter with Fastracehorse@DRF. I can honestly
say the encounter has changed my view of the handicapping process.

I]d like to thank you publically for making such a difference in my
handicapping life and in such a short time. The impact you've had is
profound.

Why have I changed? It is because of statements like these:

Originally Posted by Fastracehorse@DRF
......making #'s ain't that much work and the variant really isn't a
problem.

fffastt

Originally Posted by Fastracehorse@DRF
..........in Canada we measure the speed of a racehorse using the metric system:
METERS PER SECOND SQUARED

fffastt

Originally Posted by Fastracehorse@DRF
2nd: Figs uncover longshots because they need to b adjusted.

fffastt


Originally Posted by Fastracehorse@DRF
...........................and I hate odds lines; they aren't relevant to me.
How bout U and me do selections at Belmont on Sundays for the rest of the meet.
I don't want it to be too time consuming.

fffastt

Originally Posted by Fastracehorse@DRF
how do pace bets if U don't make selections? ( accidental pun )
I use speed figs too, not CJ's, but I adjust Beyers.
Let's do the Sunday Belmont thing - and when I beat you I promise U I'll teach
U something And I will beat U, probably pretty damn bad

fffastt

Originally Posted by Fastracehorse@DRF
......................it has to be a public lashing
it'll b good 4 u as far as furthering your game goes, and mine

fffastt

Originally Posted by Fastracehorse@DRF
My best tip
never ever ever bet a longshot
they win at such low probabiity

J/K

I swear I read the above in a handicapping book once

fffastt

As I sit at my desk today, looking at the Belmont card, and
reflecting on the advice given by Fastracehorse@DRF, I have decided
not to wager today. Instead I am embarking on several studies.

I will study the Beyer Speed Rating.
I will study how speed is proportional to sec squared.
I will study the futility of oddslines.
I will study how to adjust Beyer Speed Ratings.
I will study how longshots win at win at such low probabiity.
I will study how to make Beyer type pace (I list this study last,
because I first have to complete the study on how speed is
proportional to sec squared. This may not seem as easy as it first
appears)
Most importantly, I will not make another wager until I complete
my studies. Not another wager until I rise to Fastracehorse@DRF's
skill level.

Once again, thank you, Fastracehorse@DRF, for pointing me in the
right direction. I know with your advice and direction there is a
profit in my future.

Mike

...............at Belmont is out of the question?????????????????

lol

fffastt

[delayjf]

I was one that was chagrined when TSN went away and merged with Bris. I found their speed /pace figures usable. Not that they were in CJ or Cramers league. My first foray into using bris numbers was a disaster, when I did compare them with the Beyers they often conflicted.

On their own scale, TSN figures seemed to parallel the Beyers number more closely than the Bris numbers did. I recall CJ did a more statistical comparison between the two and concluded the same thing.

What I don't get is that both TSN and bris claimed that their numbers were made using basically the same method - so why the difference?

[CincyHorseplayer]

Can we please get Timeform Figures loadable into Excel?CJ's fig are universally the best and most respected but you can't do anything with them.Some of us like to make our own compound pace ratings from our own thoughts.It's not an insult.CJ,accurate numbers are the launching point.You are it man.

[crusader222]

Hi, been awhile.I found a new sheet rating method called WINNERMETRICS and wondered if anyone has had any experience with them.Thinking of trying them.