In search of a better track variant

[jeebus1083]

People here have posted how in figure-making, the track variant step is "part science, part art". Andy Beyer has maintained for years that projecting final time figures against the raw figures produce the best track variants man can devise. To me, that's a hollow answer. Projections are sensitive to the perceived improvement or decline of both the horse and the competition it faces. How can a track variant be based on such perception?

At 6F, the winner runs a 100 Beyer raw figure, beating his adversary by 2 lengths (95). However, the winner has A) never run anything better than a 92, and B) the runner-up's recent bests were 87-88-88. Based on these facts, a figuremaker would project the winner to have run to his best (92, or Fast 8), while the runner-up would end up with an 87, putting the efforts in line with their capabilities.

How is creating a figure that is "in line" with past performance a "scientific" or "artful" practice? Why couldn't have the winner and runner-up improved here from their previous efforts? In other words, how can one make a case for these runners to earn the 100 and 95 figures, instead of being projected downward to "fit the profile" they have drawn up over time?

Same problem exists when the winner runs a lowish figure. If the horse runs a 90, but runs between 98-102 on his best days, and the runner-up usually runs in the neighborhood of 95-97, is it prudent to project a higher figure to put the race "in line" or did the horses actually regress?

Horses have good days and bad days. I guess my question is, how can the inherent "speed" of a racetrack on any given day, be judged based on how the horses "should" be running? Is there a better way out there to gauge the condition of a racetrack other than using the horses as a tool?

I don't deny that variances in track surfaces exist day-to-day. It would be misleading to say otherwise. Weather, track maintenance, and constant running on the surfaces change how the surface behaves. However, is the degree of variance overrated? Do most track surfaces play fair on "normal" days?

[Robert Goren]

However, is the degree of variance overrated? Do most track surfaces play fair on "normal" days?No to the first. There is no such thing as a "normal" day. After a good rain, tracks very often change permanently and do not revert back to the same Variant as it was before. Tracks are constantly evolving. At least that has been my experience. Although I will give you this much, you can read too much into very small differences in adjusted speed ratings. But the real problems begin when you have no idea what your Variant is for that day. There are some day the numbers are so all over the place you have no idea what it is. At least if you make your own numbers, you what day that is.

[Pell Mell]

People here have posted how in figure-making, the track variant step is "part science, part art". Andy Beyer has maintained for years that projecting final time figures against the raw figures produce the best track variants man can devise. To me, that's a hollow answer. Projections are sensitive to the perceived improvement or decline of both the horse and the competition it faces. How can a track variant be based on such perception?

At 6F, the winner runs a 100 Beyer raw figure, beating his adversary by 2 lengths (95). However, the winner has A) never run anything better than a 92, and B) the runner-up's recent bests were 87-88-88. Based on these facts, a figuremaker would project the winner to have run to his best (92, or Fast 8), while the runner-up would end up with an 87, putting the efforts in line with their capabilities.

How is creating a figure that is "in line" with past performance a "scientific" or "artful" practice? Why couldn't have the winner and runner-up improved here from their previous efforts? In other words, how can one make a case for these runners to earn the 100 and 95 figures, instead of being projected downward to "fit the profile" they have drawn up over time?

Same problem exists when the winner runs a lowish figure. If the horse runs a 90, but runs between 98-102 on his best days, and the runner-up usually runs in the neighborhood of 95-97, is it prudent to project a higher figure to put the race "in line" or did the horses actually regress?

Horses have good days and bad days. I guess my question is, how can the inherent "speed" of a racetrack on any given day, be judged based on how the horses "should" be running? Is there a better way out there to gauge the condition of a racetrack other than using the horses as a tool?

I don't deny that variances in track surfaces exist day-to-day. It would be misleading to say otherwise. Weather, track maintenance, and constant running on the surfaces change how the surface behaves. However, is the degree of variance overrated? Do most track surfaces play fair on "normal" days?


IMO track variant is way overrated. I started working on track variant over 50 yrs ago, before Beyer's was ever heard of. This is what I deduced; if one is looking for the actual track condition and how it affects how fast or slow races may be run due solely to the condition of the track, par times of the final times are totally unreliable.
Final times of races are dependent upon the class of horses, the pace of the races and TRACK CONDITION. How does one discern which of these factors contributed most to the final time of any given race?
You may work out an average par time for different classes but to do that you must have a through knowledge of race conditions and how each secretary writes the conditions.
Regardless of class, the pace of the race still determines how fast a race will be run. The pace is determined by the class of horses, how the jockeys rate their horses and the track condition.
The biggest variable is found in the pace of the race which is probably why there have been so many books written concerning ways to interpret pace. There is just no way to tell what the pace is going to be because, as we have just seen over the weekend, until the race unfolds, even the participants are clueless as to where they are going to be or how the race will develop.
My final conclusion was that the only factor they could be considered somewhat reliable would be the track condition and how much or how little it changed from day to day. But how to do that? I always thought that using final times to judge track condition was too faulty due to the influence of pace. In the end I decided that to eliminate the pace factor I would only consider the portion of the race where the horses were running as hard as they could and no rating was involved. This could only be the stretch run or the last 1/4 mile.
I then proceeded to get par times for the last 1/4 from all the different classes running at the track. This was before winter racing and simulcasting and I was only betting the NJ tracks and Philly.
I found that there was not that much of a spread between the best and worst horses. For instance, a cheap maiden could run the last 1/4 damn near as fast as a middle claimer, providing he ran slow enough early. What I was looking for was just how fast could a last 1/4 be run on any given day. Doing this gave me the best track variant I ever used. Tracks just don't change that much normally.
I long ago got away from using figures in my handicapping, I just glance at them now, because I found much more profitable methods. I only spot play by using a combination of form analysis, proper placement as to class and distance and trainer and jockey moves. I have a limited number of clients who subscribe to my selections and some have been with me for 5 yrs or more so I must be doing something right.

http://backstretchtalks.blogspot.com/

[thaskalos]

People here have posted how in figure-making, the track variant step is "part science, part art". Andy Beyer has maintained for years that projecting final time figures against the raw figures produce the best track variants man can devise. To me, that's a hollow answer. Projections are sensitive to the perceived improvement or decline of both the horse and the competition it faces. How can a track variant be based on such perception?

At 6F, the winner runs a 100 Beyer raw figure, beating his adversary by 2 lengths (95). However, the winner has A) never run anything better than a 92, and B) the runner-up's recent bests were 87-88-88. Based on these facts, a figuremaker would project the winner to have run to his best (92, or Fast 8), while the runner-up would end up with an 87, putting the efforts in line with their capabilities.

How is creating a figure that is "in line" with past performance a "scientific" or "artful" practice? Why couldn't have the winner and runner-up improved here from their previous efforts? In other words, how can one make a case for these runners to earn the 100 and 95 figures, instead of being projected downward to "fit the profile" they have drawn up over time?

Same problem exists when the winner runs a lowish figure. If the horse runs a 90, but runs between 98-102 on his best days, and the runner-up usually runs in the neighborhood of 95-97, is it prudent to project a higher figure to put the race "in line" or did the horses actually regress?

Horses have good days and bad days. I guess my question is, how can the inherent "speed" of a racetrack on any given day, be judged based on how the horses "should" be running? Is there a better way out there to gauge the condition of a racetrack other than using the horses as a tool?

I don't deny that variances in track surfaces exist day-to-day. It would be misleading to say otherwise. Weather, track maintenance, and constant running on the surfaces change how the surface behaves. However, is the degree of variance overrated? Do most track surfaces play fair on "normal" days? It has been a while since I read Andy Beyer's books, and I could be wrong...but I don't think that he uses the "projection" method for every variant he makes.

I believe that this "art" comes into play when there is one or two "abnormal" races...which contradict what the remaining of the day's races indicate about what the variant should be.

[jotb]

IMO track variant is way overrated. I started working on track variant over 50 yrs ago, before Beyer's was ever heard of. This is what I deduced; if one is looking for the actual track condition and how it affects how fast or slow races may be run due solely to the condition of the track, par times of the final times are totally unreliable.
Final times of races are dependent upon the class of horses, the pace of the races and TRACK CONDITION. How does one discern which of these factors contributed most to the final time of any given race?
You may work out an average par time for different classes but to do that you must have a through knowledge of race conditions and how each secretary writes the conditions.
Regardless of class, the pace of the race still determines how fast a race will be run. The pace is determined by the class of horses, how the jockeys rate their horses and the track condition.
The biggest variable is found in the pace of the race which is probably why there have been so many books written concerning ways to interpret pace. There is just no way to tell what the pace is going to be because, as we have just seen over the weekend, until the race unfolds, even the participants are clueless as to where they are going to be or how the race will develop.
My final conclusion was that the only factor they could be considered somewhat reliable would be the track condition and how much or how little it changed from day to day. But how to do that? I always thought that using final times to judge track condition was too faulty due to the influence of pace. In the end I decided that to eliminate the pace factor I would only consider the portion of the race where the horses were running as hard as they could and no rating was involved. This could only be the stretch run or the last 1/4 mile.
I then proceeded to get par times for the last 1/4 from all the different classes running at the track. This was before winter racing and simulcasting and I was only betting the NJ tracks and Philly.
I found that there was not that much of a spread between the best and worst horses. For instance, a cheap maiden could run the last 1/4 damn near as fast as a middle claimer, providing he ran slow enough early. What I was looking for was just how fast could a last 1/4 be run on any given day. Doing this gave me the best track variant I ever used. Tracks just don't change that much normally.
I long ago got away from using figures in my handicapping, I just glance at them now, because I found much more profitable methods. I only spot play by using a combination of form analysis, proper placement as to class and distance and trainer and jockey moves. I have a limited number of clients who subscribe to my selections and some have been with me for 5 yrs or more so I must be doing something right.

http://backstretchtalks.blogspot.com/

Let me ask you a ?...Do you include run up time into the variant..According to the Equibase result charts from 2009 (I believe) they show a run up time..The run up time has changed since 2009 at certain distances..Do you take this into consideration?

[cj]

It has been a while since I read Andy Beyer's books, and I could be wrong...but I don't think that he uses the "projection" method for every variant he makes.

I believe that this "art" comes into play when there is one or two "abnormal" races...which contradict what the remaining of the day's races indicate about what the variant should be.

I believe he does use "projection" for most of the races. This is how he described how fast he thought the race would be run in place of using "pars".

I think what you are describing is more what I would call "breaking out" a race, or using a variant different from from other races run the same day at similar distances.

[cj]

Let me ask you a ?...Do you include run up time into the variant..According to the Equibase result charts from 2009 (I believe) they show a run up time..The run up time has changed since 2009 at certain distances..Do you take this into consideration?

I can say I have been very surprised in my studies. Changes in run up don't have nearly as large of an effect as I would have thought.

[jeebus1083]

I think that the late William Scott may have some logic when he promotes his "artificial variant" approach in the book "Total Victory at the Track". He starts with a concept of an "average" or "normal" range of DRF TVs, and any TV within that range equates to there being no adjustment to the time of the race. From there, he works off a concept of 50% reliablity, meaning that for every 2 variant points = 1 length. If a "normal" track is 18, 19, or 20, there is no adjustment. However, if the TV is 21 or 22, the track is Slow 1; 23 or 24, Slow 2; etc. The process is the same on the flip side (fast), with the wrinkle that if the TV is greater than 4 TV points fast (or Fast 2), each TV point is treated as a full length (a TV of 10 with a "normal" track range of 18-20 would be Fast 6; Fast 2 until you get to 14, then Fast 4 until 10). Scott's rationale for this adjustment is that anytime the DRF TV is greater than 4 TV points fast, the track is becoming super-speedy, and should be treated accordingly.

Let me state that I'm not advocating the use of DRF's TV. I do however, feel that the premise that there is a "normal" range for track resiliency, as even the best par times out there cannot realistically be 100% accurate. Tracks change day in and day out, so the expected times will not always correlate to reality. If one could set a "margin of error range" for the track pars (in Horsestreet, it is 100) at each racetrack, it may go a long way in building better variants.

I hope I'm being clear. If anyone is lost, I'll try and clarify my position better.

[BIG49010]

I can say I have been very surprised in my studies. Changes in run up don't have nearly as large of an effect as I would have thought.

I've found the same thing, and makes me wonder if they are just a number they put in the charts just to make some people happy.

[cj]

I've found the same thing, and makes me wonder if they are just a number they put in the charts just to make some people happy.

I've watched quite a few replays and most tracks are accurate. I won't say all, but most do a good job.

[jeebus1083]

I believe he does use "projection" for most of the races. This is how he described how fast he thought the race would be run in place of using "pars".

I think what you are describing is more what I would call "breaking out" a race, or using a variant different from from other races run the same day at similar distances.

Which leads me to my next question: if tracks can change so much throughout the course of the day, wouldn't it make more sense to for the variant to be based on single races rather than the entire day, or particular periods throughout the day?

Jerry Brown at Thorograph has an informative (and dry) PowerPoint presentation on changing track conditions from the '04 Expo on his website.

[cj]

Which leads me to my next question: if tracks can change so much throughout the course of the day, wouldn't it make more sense to for the variant to be based on single races rather than the entire day, or particular periods throughout the day?

Jerry Brown at Thorograph has an informative (and dry) PowerPoint presentation on changing track conditions from the '04 Expo on his website.

They can change a lot throughout the day, but I wouldn't say this is the norm. I've seen the presentation, and while some is good, some clearly has a personal agenda in my opinion.

[jeebus1083]

They can change a lot throughout the day, but I wouldn't say this is the norm. I've seen the presentation, and while some is good, some clearly has a personal agenda in my opinion.

I agree 100% there, as he knocks Len Ragozin's methodologies more than once in the presentation. Nevertheless, Brown has a point. Track maintenance of some sort will always have some sort of effect on the surface, even if it's minute. The surface run on at 1:00, might not be the surface run on at 2:00, even if the track is labeled "fast". Some tracks will run two divisions of the same race due to overflow, and the times of those races may vary greatly, enough where people will say that Group A was vastly superior to Group B. This might be true, but if the track was maintained a certain way between those races, that might have had some effect on the final time, even if the horses in Group A were indeed a little faster.

[jotb]

I can say I have been very surprised in my studies. Changes in run up don't have nearly as large of an effect as I would have thought.

Probably not at every distance but then again it's possible it could make a difference at another distance..For example in 2009 according to Equibase on Jan 1st the run up at a 1/70 was 48ft for a given race..In 2010 on the same date the run up was 77FT, 73 FT and 74 FT..In 2011 on the same date it was 60FT, 65 FT and 66FT..Wouldn't you think from 48FT to 77FT would make some type of difference? Don't you think a horse would be traveling at a faster velocity by the time he or she reaches the timer at 77FT vs 48FT?

[thaskalos]

I believe he does use "projection" for most of the races. This is how he described how fast he thought the race would be run in place of using "pars".

I think what you are describing is more what I would call "breaking out" a race, or using a variant different from from other races run the same day at similar distances. CJ,

In your opinion...does the presence of today's "supertrainers" - and their miraculous "ways" - complicate the task of track variant projection?

[jeebus1083]

CJ,

In your opinion...does the presence of today's "supertrainers" - and their miraculous "ways" - complicate the task of track variant projection?

I know that the question was intended for CJ, but I'll chime in anyway. Horses that run big first time for a supertrainer, usually will run at a level that they have never eclipsed before. Using a projection approach, if the horses behind the juiced-up winner run their "normal" race, then the fig is probably legit. However, if the same horses run races well beyond their norm, then I could see where a projection would become a headache, as you risk producing a figure that may not be "trustworthy" until proven otherwise (i.e. next start).

[gm10]

I can say I have been very surprised in my studies. Changes in run up don't have nearly as large of an effect as I would have thought.

I would guess that there is a small but not insignificant effect, but it is probably hard to measure due to other variables adding a lot of noise to the signal. Have you tried focusing on days with minimal variation (=all races were run at a true pace, no unusual distances, stable weather, no track maintenance performed in between races, etc)?

[cj]

I would guess that there is a small but not insignificant effect, but it is probably hard to measure due to other variables adding a lot of noise to the signal. Have you tried focusing on days with minimal variation (=all races were run at a true pace, no unusual distances, stable weather, no track maintenance performed in between races, etc)?

Sure, there is an effect. However, runup at most tracks stays very constant. When it changes by a few feet, there is basically no change.

From track to track there are big differences, but these have been built in by sharp handicapper's for a long time now.

[cj]

CJ,

In your opinion...does the presence of today's "supertrainers" - and their miraculous "ways" - complicate the task of track variant projection?

It makes it tougher, sure, but not unmanageable. I would say many tend to underrate the performances of their new horses, leading to overlays when the also rans run back.

[cj]

Probably not at every distance but then again it's possible it could make a difference at another distance..For example in 2009 according to Equibase on Jan 1st the run up at a 1/70 was 48ft for a given race..In 2010 on the same date the run up was 77FT, 73 FT and 74 FT..In 2011 on the same date it was 60FT, 65 FT and 66FT..Wouldn't you think from 48FT to 77FT would make some type of difference? Don't you think a horse would be traveling at a faster velocity by the time he or she reaches the timer at 77FT vs 48FT?

I'm not saying there is no difference, just that it is minimal. That is definitely true for the near 30 foot difference you cite.

I will say changing the run up for 0 to 30 is a LOT different than changing it from 50 to 80.

[jotb]

I'm not saying there is no difference, just that it is minimal. That is definitely true for the near 30 foot difference you cite.

I will say changing the run up for 0 to 30 is a LOT different than changing it from 50 to 80.

I certainly agree with that...

[RXB]

I will give you this much, you can read too much into very small differences in adjusted speed ratings. But the real problems begin when you have no idea what your Variant is for that day. There are some day the numbers are so all over the place you have no idea what it is. At least if you make your own numbers, you what day that is.

Some very good points here.

People might be amazed at the frequency and degree of variances among the individual race figures of various practitioners. And I'm talking about guys who are all legit, who know what they're doing and who are fairly comparable in terms of overall accuracy. When their respective numbers are closest together, that's when the numbers are most reliable. When their numbers are relatively scattered compared to each other, take them with multiple grains of salt instead of just a single grain.

Having a mastery of the "art" will improve results over a purely "scientific" algorithm. Yet, the more that you can depend on the "science" and the less than you need to resort to the "art," the more reliable your figures will be. As soon as you start breaking out races or making multiple variants for similar distances within a day, it's a red flag in terms of reliability.

I would take last year's Ky Derby as a classic example. More than two hours between dirt races on a muddy but drying-out track; the first 6f much faster relatively than the final time, and a raw final time that would've yielded an impossibly slow depiction of the real calibre of the race. Talk about variables to introduce error potential into a speed figure! Ragozin has always argued for a single daily variant but in this case how could anyone possibly believe that the track hadn't slowed considerably? The question was, how considerably? The Beyer boys and I estimated it exactly the same; CJ was several lengths less lenient in terms of how much the track had slowed. Who's to say whether one estimate was better than the other, or if so, by how much? In those situations, you do the best that you can do, but the most important thing is to recognize that you cannot possibly invest much confidence in whatever number you've assigned.

Cases as drastic as that are rather infrequent but there is some degree of error potential for every figure. And as RG pointed out, the more that you aware of this, the better off you will be.

[cj]

I would take last year's Ky Derby as a classic example. More than two hours between dirt races on a muddy but drying-out track; the first 6f much faster relatively than the final time, and a raw final time that would've yielded an impossibly slow depiction of the real calibre of the race. Talk about variables to introduce error potential into a speed figure! Ragozin has always argued for a single daily variant but in this case how could anyone possibly believe that the track hadn't slowed considerably? The question was, how considerably? The Beyer boys and I estimated it exactly the same; CJ was several lengths less lenient in terms of how much the track had slowed. Who's to say whether one estimate was better than the other, or if so, by how much? In those situations, you do the best that you can do, but the most important thing is to recognize that you cannot possibly invest much confidence in whatever number you've assigned.



Just a couple points about this, and please, realize I'm not taking anything you wrote as criticism.

I use pace calls in addition to final time, and I think this gives me a better picture of how the track might have changed. Also, I monitor races where I have questions. In the case of the Derby, nothing even came close to making me think I was wrong for the rest of the season. Even that doesn't mean I was right, but the figure never hurt me at the betting windows.

That doesn't, however, mean I don't get any wrong...I do, and we all do.

[Broad Brush]

People here have posted how in figure-making, the track variant step is "part science, part art". Andy Beyer has maintained for years that projecting final time figures against the raw figures produce the best track variants man can devise. To me, that's a hollow answer. Projections are sensitive to the perceived improvement or decline of both the horse and the competition it faces. How can a track variant be based on such perception?

At 6F, the winner runs a 100 Beyer raw figure, beating his adversary by 2 lengths (95). However, the winner has A) never run anything better than a 92, and B) the runner-up's recent bests were 87-88-88. Based on these facts, a figuremaker would project the winner to have run to his best (92, or Fast 8), while the runner-up would end up with an 87, putting the efforts in line with their capabilities.

How is creating a figure that is "in line" with past performance a "scientific" or "artful" practice? Why couldn't have the winner and runner-up improved here from their previous efforts? In other words, how can one make a case for these runners to earn the 100 and 95 figures, instead of being projected downward to "fit the profile" they have drawn up over time?

Same problem exists when the winner runs a lowish figure. If the horse runs a 90, but runs between 98-102 on his best days, and the runner-up usually runs in the neighborhood of 95-97, is it prudent to project a higher figure to put the race "in line" or did the horses actually regress?

Horses have good days and bad days. I guess my question is, how can the inherent "speed" of a racetrack on any given day, be judged based on how the horses "should" be running? Is there a better way out there to gauge the condition of a racetrack other than using the horses as a tool?

I don't deny that variances in track surfaces exist day-to-day. It would be misleading to say otherwise. Weather, track maintenance, and constant running on the surfaces change how the surface behaves. However, is the degree of variance overrated? Do most track surfaces play fair on "normal" days?

To answer your original question: Is there a better way create a variant?

I would say no. Many have pointed out the possible errors that can happen in
calculating a variant based on what horses should run. However, I see no other way to determine what it should have been. If you could precisely measure a racing surface to know how fast or slow it was--this would not work to calculate a variant. The surface could be exactly the same on two days but
the temperature, humidity and head winds can cause different variants.
You can see this in Track & Field with human runners---the surface itself is the same--but a different day causes different times. People get mad when Beyer
goes back and changes a Fig later. Sometimes this is the only way because there was not enough data to compare to and come up with an accurate Fig.

[Tom]

I suppose use of Trakus could improve the process.

[therussmeister]

Some very good points here.

People might be amazed at the frequency and degree of variances among the individual race figures of various practitioners. And I'm talking about guys who are all legit, who know what they're doing and who are fairly comparable in terms of overall accuracy.

This is one of the two reasons why I make my own figures. Often enough my figures show as much as an eight length difference compared to Beyer's (the only figures I can compare mine with), and sometimes my figures are right - at a big price. When they are wrong? Oh well - I miss a chance at cashing at 9/5.

The other reason is, if you make your own, you can attach a level of confidence to each days figures.

[Cratos]

People here have posted how in figure-making, the track variant step is "part science, part art". Andy Beyer has maintained for years that projecting final time figures against the raw figures produce the best track variants man can devise. To me, that's a hollow answer. Projections are sensitive to the perceived improvement or decline of both the horse and the competition it faces. How can a track variant be based on such perception?

At 6F, the winner runs a 100 Beyer raw figure, beating his adversary by 2 lengths (95). However, the winner has A) never run anything better than a 92, and B) the runner-up's recent bests were 87-88-88. Based on these facts, a figuremaker would project the winner to have run to his best (92, or Fast 8), while the runner-up would end up with an 87, putting the efforts in line with their capabilities.

How is creating a figure that is "in line" with past performance a "scientific" or "artful" practice? Why couldn't have the winner and runner-up improved here from their previous efforts? In other words, how can one make a case for these runners to earn the 100 and 95 figures, instead of being projected downward to "fit the profile" they have drawn up over time?

Same problem exists when the winner runs a lowish figure. If the horse runs a 90, but runs between 98-102 on his best days, and the runner-up usually runs in the neighborhood of 95-97, is it prudent to project a higher figure to put the race "in line" or did the horses actually regress?

Horses have good days and bad days. I guess my question is, how can the inherent "speed" of a racetrack on any given day, be judged based on how the horses "should" be running? Is there a better way out there to gauge the condition of a racetrack other than using the horses as a tool?

I don't deny that variances in track surfaces exist day-to-day. It would be misleading to say otherwise. Weather, track maintenance, and constant running on the surfaces change how the surface behaves. However, is the degree of variance overrated? Do most track surfaces play fair on "normal" days?

First and foremost to attempt to calculate a unique variant from two things that are both varying (the horse and the racetrack surface) is virtually statistically impossible.

To solve any problem mathematically with a single answer there must be a constant in the equation. If everything varies there is no unique answer.

Therefore a horse standing in the gate has potential energy and once the race begins it exhibits kinetic energy. That kinetic energy in perfect form yields maximum effort by the horse. However given that potential energy is cyclical and varies from one horse outing to another there isn’t any way to determine what kinetic energy output will be yielded by the horse in any given race.

On the other hand, maintenance, environmental conditions, and track use can dramatically change the track surface from race to race and the most salient environmental effect will be the effect of wind.

The effect of run-up is acceleration. The longer the run-up; the greater the velocity and produces a greater acceleration. But the effect of that velocity on final time depends on where the race starts in respect to the turn.

With the same run-up distance, Belmont with its over 2,000 feet turns would have a different final time affect than Saratoga with its nearly 1,500 feet turns

[RXB]

I use pace calls in addition to final time, and I think this gives me a better picture of how the track might have changed. Also, I monitor races where I have questions. In the case of the Derby, nothing even came close to making me think I was wrong for the rest of the season. Even that doesn't mean I was right, but the figure never hurt me at the betting windows.


http://paceadvantage.com/forum/showthread.php?t=69764&page=4&pp=15

I found the thread about last year's Derby figures. I thought it was interesting on a number of levels. I've linked to the page where we were sharing our reasoning for assigning the numbers that we did. A lot of factors can be taken into account. Pace is definitely something that I consider.

My phrase about being "several lengths" apart was not the best choice of words as it was only five points (~3 lengths). "A few lengths" would've been more appropriate.

[jeebus1083]

What if one could apply Scott's 50% reliability principle to a par times chart instead of DRF's TV?

At most tracks, the $10,000 claiming class is base par (or 100). However, the NW1 level is considered the class of the track, which in the conventional Quirin approach, could be for this example, a 104, depending on the structure of the track class hierarchy. What if we used the base and the NW1 as a premise for a zero adjustment range for a "fair" track? To widen the range, we would take the difference (104-100 = 4) from the higher end, and apply that downward, so the zero adjustment range would be any figure earned between 96 and 104. That's a range of 1 3/5 seconds (8 ticks).

Now that a range has been set, the same rules for adjustments in the quote box below would apply.

I want to experiment with this, and see what results can be had.

I think that the late William Scott may have some logic when he promotes his "artificial variant" approach in the book "Total Victory at the Track". He starts with a concept of an "average" or "normal" range of DRF TVs, and any TV within that range equates to there being no adjustment to the time of the race. From there, he works off a concept of 50% reliablity, meaning that for every 2 variant points = 1 length. If a "normal" track is 18, 19, or 20, there is no adjustment. However, if the TV is 21 or 22, the track is Slow 1; 23 or 24, Slow 2; etc. The process is the same on the flip side (fast), with the wrinkle that if the TV is greater than 4 TV points fast (or Fast 2), each TV point is treated as a full length (a TV of 10 with a "normal" track range of 18-20 would be Fast 6; Fast 2 until you get to 14, then Fast 4 until 10). Scott's rationale for this adjustment is that anytime the DRF TV is greater than 4 TV points fast, the track is becoming super-speedy, and should be treated accordingly.

Let me state that I'm not advocating the use of DRF's TV. I do however, feel that the premise that there is a "normal" range for track resiliency, as even the best par times out there cannot realistically be 100% accurate. Tracks change day in and day out, so the expected times will not always correlate to reality. If one could set a "margin of error range" for the track pars (in Horsestreet, it is 100) at each racetrack, it may go a long way in building better variants.

I hope I'm being clear. If anyone is lost, I'll try and clarify my position better.