what if one could figure comparable workout times?

[JohnGalt1]

[QUOTE=Fingal]Times can be very misleading. Take for example Get Happy Mister who won the San Simeon Hcp. at Santa Anita this year. Not only is he a Colorado bred, but his last work before the race while being on the turf was 6f in 1:19. Now that has to be awful close to the threshold for having a timed work. Horse paid 27 bucks.


Unless the work out is very fast, slow workouts like this can mean a 135+ lb. exercise rider, so I don't downgrade the time.

I'd rather see slow than no workouts.

[classhandicapper]

I never did a formal study like Dave, but my experience from handicapping thousands of races is similar. Improving workouts can signal a solid or improved performance, but especially for FTS and lightly raced horses.

[Capper Al]

Originally Posted by Augenj

This doesn't sound like what you're looking for but it might trigger some thought.

You can use average workout rankings back any number of workouts to get a "rating" for each horse. Let's say you want to use the last 3 workouts for your calculations. Sum up the last 3 rankings for the horse (X). Then sum up the last 3 total ranked horses (Y). Divide X by Y and subtract the result from 1.0 to get the best on top.

The win per cent over hundreds of races is low but the ROI is pretty decent. It all depends on the track and type of race - maidens, non-maidens. I use an enhanced version of this as one of 10 factors to calculate a horse's overall rating.

I don't understand this.

[Capper Al]

Originally Posted by Tom

The workout rating in HTR is one of the most powerful ones they have.
Not as a stand alone, but when combined with other factors. It is in several of my most productive spot plays.

I agree with other factors. I don't know HTR's workout ratings, but I'm guessing it probably is a conversion of feet per second at distance.

[Capper Al]

Originally Posted by pandy

Has anyone ever done any record keeping or testing off workouts? For instance, do horses with a recent work in last 7 days win more often than horses that don't have a recent workout? Do horses that have a recent bullet workout win more often?

Although it's true that if you don't see the work, you don't know if it was good or not, it still gives you an idea of what the trainer is trying to do in terms of conditioning.

Also with older horses, if the horse is working regularly it's a good sign, the horse is sound.

I have had several big scores because of workouts. You have to use common sense. As an example, some years ago Ron McAnally had a first time starter, I can't remember her name right now, but she was working fast. But his stats on firsters was something like 0 for 68. However, the works were unusually fast for a horse from his barn, a sign that the horse was indeed fast, since he doesn't work his maidens fast. I had bet the horse and gave it out and then luckily they interviewed him on TVG and Ron said that although he never wins with firsters, "I can tell you that this one can run." I doubled my bet and she won easily at $46.

I did study workouts before and am about again with my program rewrite.

[Capper Al]

Originally Posted by Dave Schwartz

Pandy,

I have. There is validity to recent workouts. Percentages & Probabilities, page 89.

However, the difference in IVs is pretty slight. High point is 5-7 days, and tops at only 1.11, with the bottom being no work in 30 days at 0.82. Seems like not being worth the effort.

Because some tracks demand a workout for a horse returning from layoff and others do not, the statistics are skewed unless they are track-specific.


However...

When one uses a universal parallel time chart and looks at "Best Workout in 45 Days," one finds that horses ranked 1st perform well enough to actually be profitable at high odds.


However #2...

In my handicapping experience, even that factor is "correlational rather than causational," and, in practice, consistently points one to the wrong horses.


However #3...

The exception comes in young horses, races with FTS, and lightly-raced horses, where the 45-day best workout blossoms into a factor that can actually be used to handicap.

In fact, FTS that rank #1 for BW45 and go off at (I believe it is) 12/1 or higher have shown to be flat bet profitable.

I scored some nice hits with parellel charts on workouts.

[Capper Al]

Originally Posted by classhandicapper

I never did a formal study like Dave, but my experience from handicapping thousands of races is similar. Improving workouts can signal a solid or improved performance, but especially for FTS and lightly raced horses.

Isn't this the truth about any figure in horse racing. It is never the best number or the best average number, it's how the racing game reflects that number.

[Dave Schwartz]

This doesn't sound like what you're looking for but it might trigger some thought.

You can use average workout rankings back any number of workouts to get a "rating" for each horse. Let's say you want to use the last 3 workouts for your calculations. Sum up the last 3 rankings for the horse (X). Then sum up the last 3 total ranked horses (Y). Divide X by Y and subtract the result from 1.0 to get the best on top.

The win per cent over hundreds of races is low but the ROI is pretty decent. It all depends on the track and type of race - maidens, non-maidens. I use an enhanced version of this as one of 10 factors to calculate a horse's overall rating.

Very interesting idea. Theoretically one could do this with many other things as well.

I could see scanning the back pacelines for each horse creating (say) a 3-race average and then the next race looking for patterns.

Very smart.

[Augenj]

Originally Posted by Capper Al

I don't understand this.

Me either (joking). In the past performances in the "WORKS" line, the last item on the line are 2 numbers separated by a slash, let's say 4/16. The 4 is this horse's "rank" in the number of horses (16) that had workouts that day at that distance, or "total ranked". I got these names from my data provider so sorry for the confusion. In the example:

X = the sum of this horse's "ranks" over the last 3 workouts.
Y = the sum of the "total ranked" over the last 3 workouts.
Divide Y into X to get this horse's rating to compare against other horse's ratings. Usually it will be less than 1.0.

Trouble is that the lowest decimal value is actually the best rating so to reverse it we subtract this horse's rating from 1.0 to compare against other horse's ratings. The results should always be less than or equal to 1.0 with the highest rating on top.

[Augenj]

Originally Posted by Dave Schwartz

Very interesting idea. Theoretically one could do this with many other things as well.

I could see scanning the back pacelines for each horse creating (say) a 3-race average and then the next race looking for patterns.

Very smart.

Thanks. Not smart but sometimes I make enough mistakes to learn a tiny bit.

[Capper Al]

Originally Posted by Augenj

Me either (joking). In the past performances in the "WORKS" line, the last item on the line are 2 numbers separated by a slash, let's say 4/16. The 4 is this horse's "rank" in the number of horses (16) that had workouts that day at that distance, or "total ranked". I got these names from my data provider so sorry for the confusion. In the example:

X = the sum of this horse's "ranks" over the last 3 workouts.
Y = the sum of the "total ranked" over the last 3 workouts.
Divide Y into X to get this horse's rating to compare against other horse's ratings. Usually it will be less than 1.0.

Trouble is that the lowest decimal value is actually the best rating so to reverse it we subtract this horse's rating from 1.0 to compare against other horse's ratings. The results should always be less than or equal to 1.0 with the highest rating on top.

Okay. I got it now. Thanks.

[Capper Al]

Originally Posted by Dave Schwartz

Pandy,

I have. There is validity to recent workouts. Percentages & Probabilities, page 89.

However, the difference in IVs is pretty slight. High point is 5-7 days, and tops at only 1.11, with the bottom being no work in 30 days at 0.82. Seems like not being worth the effort.

Because some tracks demand a workout for a horse returning from layoff and others do not, the statistics are skewed unless they are track-specific.


However...

When one uses a universal parallel time chart and looks at "Best Workout in 45 Days," one finds that horses ranked 1st perform well enough to actually be profitable at high odds.


However #2...

In my handicapping experience, even that factor is "correlational rather than causational," and, in practice, consistently points one to the wrong horses.


However #3...

The exception comes in young horses, races with FTS, and lightly-raced horses, where the 45-day best workout blossoms into a factor that can actually be used to handicap.

In fact, FTS that rank #1 for BW45 and go off at (I believe it is) 12/1 or higher have shown to be flat bet profitable.

Reviewed P&P 2012 section on workouts. Always a good resource.

[Dave Schwartz]

Thank you, Al.

Looking back on the factor choices in our database, they could be improved quite a bit.

I think the chapter on recency surprised me a lot with some of the changes, or at least perceived changes from the "old days."

For example, 5 "times on track" (i.e. races + works) in a month used to be a good thing. Now the stats for that 5th trip have dropped.

Another example was "furlongs raced+worked in" 14 days, 21 days, 28 days. Used to be considered a good thing if a horse raced and/or worked a furlong a day. Not so any more.

[Tom]

Originally Posted by Capper Al

I agree with other factors. I don't know HTR's workout ratings, but I'm guessing it probably is a conversion of feet per second at distance.

No, more than just simple speed.

[Capper Al]

Originally Posted by Tom

No, more than just simple speed.

Funny, more than speed is going into my rewrite numbers also. I could be reinventing the wheel.