Speed figure study

[Tom]

Doing a lot of studying in speed figure patterns, bounces, recovery lines, etc. I found an intersting factoid u sing Crammer speed numbers, and I assume Beyers would have a similar situatuation. The top fig horse off it's last race won 26% of the time. But if that last out speed figure equalled the hors'e best fig in the last 90 days, the win % dropped to 16%. Good elimination factor. I am going to do some deeper querys alon this line, but this seems to me to have similar value to Hurrikane's morning line favorites thorw outs he posted a while back - if you can thorw out low proiced horses or top fig horses, the value of the race opens up. I can only check this angle with Crammer numbers and 90 day time spans becasue that is what my db has in it. Would be interesting to see if anyone can look at this with other figs and other time frames.

[Zaf]

Thats interesting and useful information. Thanks for posting TOM !

ZAFONIC

[cj]

Tom,

You say equalled, can I assume you mean equalled or surpassed? Probably a stupid question, but that is some valuable stuff, want to make sure I understand it perfectly before experimenting with it with my homemade figs as well as Beyers.

Thanks in advance

[hurrikane]

yeah, I read that Tom. Very interesting.

By the way...the MLO thing was

19% rMLO =1 rKrat >2

13% rMLO = 1 rKrat >4

works well looking for open races and exotics.

[turfspec]

Tom,

When I see a result like that my question is why? Do you think it's because the majority of horses sharing that trait decline off of two tops in the 90 day period or could it be due to some other factor like the trainer moving the horse up in class off a top race? I think the reason for the decline in win% can change the dynamic and possibly point out a profitable subset among those that do manage to win.

Rob

[hurrikane]

good points turf.

It would also behove you to break it out by class grps. Stk Alw Clm Mdn MCL.

The thinking is it likley represents a peak. And chances are that peak is more detrimental to a lower quality horse than a higher quality horse.

[Valuist]

That is interesting. I wonder what the ROI would've been.

I did a study of Beyer numbers back in the mid 90s and found strong evidence that the horse with the best figure in its LAST 10 RACES (at similar distances and on fast or good trcks) far outperformed best Beyers in their last race. The best in L10 races group had positive ROIs in both routes and sprints.

[alysheba88]

Just remember todays winner on average, improved his Beyer around 7 points out last time. Having the last best Beyer does not mean much.

[andicap]

Interesting research project:

--Last Best Beyer (Cramer/Colts Neck/BRIS, etc.)
--Under 6 or 9 month top by say at least 7 points (3-4 lengths -- picked at random, optimum figure couldb e different)
-- Throw out any race where the top is on a track other than fast/firm

break into two subsets
- horses coming back within 21 days ("hot" horses)
-- Horses off 21-45 days. (bit fresher)

(so you don't rate top Beyers off of layoffs)

I would also then break out competent trainers from those that rarely win.

I would break out 2-3 yr olds from older horses since younger horses tend to recovery more quickly and even throw 2 or 3 tops together (like War Emblem and Funny Cide did) before collapsing.

-----
I would also be interested in a similar project with a horse who has the best speed rating two races back as long as it's within 3 months and was on a fast/firm track. Caveat would be they did not equal or surpass that rating in last race.

[Tom]

My data has a filed for last race crammer , and C90-best Crammer in 90 days. So I set my querry to last race crammer = C90. If it is the best or an improvment, I cannot tell from what I have. Hurrickne-I run those querry's every day and look at those races that might offer value first. I simply throw out the mlo1 horse and if he beats me, he beats me, but I have a better chance of profits when I can find these races. I will scratch the horse , then use HTR as if it were never there.

[Dave Schwartz]

Crammer, huh?

I'd wouldn't call him that to his face, if I were you. <G>

Dave

[GR1@HTR]

Thanks for sharing Tom. I had KM create an export for me after reading Blinkers Off so I could test pace tops and final time tops along w/ a few other things in the book.

[Tom]

Originally posted by turfspec
Tom,

When I see a result like that my question is why? Do you think it's because the majority of horses sharing that trait decline off of two tops in the 90 day period or could it be due to some other factor like the trainer moving the horse up in class off a top race? I think the reason for the decline in win% can change the dynamic and possibly point out a profitable subset among those that do manage to win.

Rob

....was that the ones coming back in 10-20 days would be big losers and those waitning 30-45 days would be bigger winners. Not so. Days didn't matter in my fist look see. I need to look deeper into these horses this weekend. I think there might be something to adding in a factor or two to this subset and see what shakes out.

[Tom]

Dave,
Of course I meant Kramer <G>

Glen,
Long time no post - how's the new job going?

[GR1@HTR]

Tom,

New jobs sucks. It's taking too much time away handicapping...But that will change soon...

[hurrikane]

Gr1.

looks like the pup has inherited some of your social graces.

Too bad about the job. Hope you can find a cure real soon.

[delayjf]

Interesting that you did that, I to read the book and was thinking of doing the same thing. I believe that cramers pace figs would mirror the Xtras in terms of energy distribution.

So what have you found so far??

[pmd62ndst]

Here's what I found using Beyer Speed Figures:

Scenario #1: Horse with the highest last speed figure:

ROI (%) Wins/Races (%) Revenue/Wins
BBLast Rank = 1
$ -8670.80 (-14.5%) 8304/30000 (27.7%) $51329.20/8304 ($6.18)


Scenario #2: Horse with the highest last speed figure that was also the horse's highest speed figure in the last 90 days of the race.

Beyer Best Last Rank = 1
Beyer Best Last = Beyer Best 3Months
$ -5908.80 (-13.8%) 5859/21419 (27.4%) $36929.20/5859 ($6.30)


Scenario #3: Horse with the highest last speed figure that had a higher speed figure within the last 90 days of the race.

Beyer Best Last Rank = 1
Beyer Best Last < Beyer Best 3Months
$ -1591.00 (-12.8%) 1941/ 6194 (31.3%) $10797.00/1941 ($5.56)


Scenario #4: Horse with the highest last speed figure but hadn't raced in the last 3 months.

Beyer Best Last Rank = 1
Beyer Best 3Months = null
$ -1171.00 (-24.5%) 504/ 2387 (21.1%) $3603.00/504 ($7.14)



OK, so not exactly the 16% win rate that Tom stated.

Scenario #2,3 and 4 are subsets of #1, meaning #1 = #2+#3+#4.

Although Tom's theory is based on Scenario #2, I think that Scenario #3 and #4 offer a more significant angle. Scenario #3 has an increase in both ROI% and Win % while Scenario #4 shows a decrease in both ROI% and Win %.

Will research further.

PMD

[Fastracehorse]

It's very difficult to be certain about bounces after top figs.

And, the opposite is true.

I have noticed that once a trainer gets a horse good: The horse can be and often is, very dangerous. I call it a good-2 angle.

If they win at a 16% clip as Tom may suggest - that can still ruin a few good days.

fffastt

[andicap]

the point isn't whether a few can run good. The question is whether the few that run good are overbet and a greater proportion of them don't run well.

Drugs have likely changed matters since Ragozin's salad days so the bounce isn't as major a factor as it had been years ago, but if high Beyers are overbet in relation to their chances, that's significant.

Also, not sure if best Beyer in last 90 days is the best barometer. That may still not represent a real top. Lots of horses run 90-day tops that are still well below their tops from the past year. And what about horses that run
90-day tops because they only have 1 or 2 races during those 90 days? Anything they would run would be a top.

So as usual, database mining can be pretty distorted.

[pmd62ndst]

Looking back, I had to ask myself why are 2/3rds of all horses with a top last Beyer figure get it in the last 3 months? That does seem unusually high and that's because for a majority of them, they only ran 1 or 2 races in the last 3 months.

[pmd62ndst]

Example:

Maiden Special Weight card that has all first-time starters except one horse that ran a 20 Beyer two months ago and a 10 Beyer a month ago.

This horse will fall under my Scenario #3 where the horse had the highest last speed figure and had a higher speed figure within the last 90 days of the race.

I wouldn't place a bet on this horse with my worst enemy's money.

PMD

[Fastracehorse]

the point isn't whether a few can run good. The question is whether the few that run good are overbet and a greater proportion of them don't run well.

===============================

16 % is a dangerous win clip, regardless of the odds. Your question is not my question Andi - my question is this: "Am I going to sacrifice a p-3 ticket based on a possible bounce."

Also, the ROI on these types may be better than you think. Especailly if the horse steps up in class after a good effort - he will be 'less' overbet.

fffastt

[Fastracehorse]

Looking back, I had to ask myself why are 2/3rds of all horses with a top last Beyer figure get it in the last 3 months

++++++++++++++++++++++++++++++++++

Could you re-phrase that please so it makes sense.

fffastt

[pmd62ndst]

>>>

Looking back, I had to ask myself why are 2/3rds of all horses with a top last Beyer figure get it in the last 3 months

>>>

*laugh* Too much coffee there.

Here's what I meant to say:

Out of 30000 horses, 21419 of them (little over 2/3) have gotten their highest Beyer Figure within the last 3 months. This is misleading because a good portion of them have only run 1 or 2 races in the last 3 months.

[Fastracehorse]

I see.

:eek: :cool: :rolleyes: :mad: :D ;) :p

fffastt

[creatureman]

all this sounds similiar to the lowest bounce point theory does it not??????

[Tom]

Thanks for sharing, pmd62ndst...interesting data.
fffast...all too true! I get beat way too often by predicted bounces, which is what lead me to my quest. Hoping to find something to guide me better. Although, even when a horse really bounces, it still might win, if its competition is weak, I think I see this phenom at cheap tracks a lot. At FL, a shipper from NYRA might win in a 78, come back and regress to a 65, and still beat bunch of 50's by daylight.
Andi...I want either a good value play or a good elimination factor. Not particular.

[hurrikane]

pmd2ndst,

This is my point for testing by class. clm, alw, stks, mdn. mcl.

these horses all race at different intervals. A clm has to pay his feed running in low paying claiming races. a stks horse may only run 2 times in 90 days.
Also by age. 2 , 3 and then older.

My feeling is this best last out 90 day piece is going to have it's most significance with older claimers. And that may be the only place it is worthwhile

[Show Me the Wire]

How about this scenario, horse wins and runs to his new top while being dropped 30% in claiming price for trainer A and is claimed out of the race. Trainer B (new connections) runs horse within 90 days increasing the claiming price by 20% and horse loses and runs a lower speed figure. Is that a bounce?

Instead, new trainer runs horse for 10 % less than the original claiming price and horse wins with an equal speed figure. Is that a bounce?

How do the above-mentioned scenarios impact the fact that 2/3 of the horses had their best speed figure in the last 90 days?

Regards.
Show Me the Wire


Perception is reality

[Fastracehorse]

I get beat way too often by predicted bounces, which is what lead me to my quest.

++++++++++++++++++++++++++++++++

I hate chalk so you know I've been beaten alot too - but no more I say!! :cool:

Quests are awesome: Because you usually find what you are looking for.

I've answered some amazing questions for myself that last couple of years - and people think there isn't much to learn :rolleyes:

fffastt

[GR1@HTR]

Delay,

I've never used the Blinkers Off sheets so I cant really comment...But for the most Part, I think pace figs are about the same 85% of the time...JMHO...

[Tom]

When I think about all that I know, I have to believe there must be more! :eek:

[delayjf]

GR1@HTR,

Me neither, but it is my understanding that the Cramer pace figs maintain the energy relationships in the same manner the Extras do. So, his theories should hold up with the cramer numbers as well. Are you finding that to be true using the Carmer pace figures?

[andicap]

Dont velocity figures -- like those used with HTR-- maintain that same relationship as well?

[delayjf]

ANDICAP,

I suspect that they do not, I e-mailed Cary Fortias who explained to me that if you "normalize" your figures or attempt to adjust for distance that the energy relationships are altered. Actually, it was Keilan, who does post here, who explained that one of the main reasons so many handicappers have problems appling % early energy is that the distance equalization adjustments alter the energy relationsips. Therefore one needs to model %early by distance (amoung other things).

Ken Massa applies a home made distance adjustment to his velocity figs that Cramer does not. Knowing Ken, this adjustment serves a valueable purpose in predicting performance at another distance, but it still affects the energy relationships.

As with the Cramer pace figures, the extras deal in track varient adjusted velocity only. He may adjust for other factors, but I'm not sure right at this moment.

[Fastracehorse]

A speed figure is less work and just as effective as pace figures, ( IMO more effective ).

I cringe at all that work.

:cool:

fffastt

[delayjf]

Fast,

Agreed, but with HTR you don't do any work, they are there for you.

[Fastracehorse]

Why don't we pick one hundered horses to bet over the next 3 months - based on pace ratings at HTR and based on adjd speed figs.

That would be fairly easy for me to do -- I don't know hom much you like to 'cap - we can compare ROI's just for fun.

Who knows - you might kill me.

fffastt

[andicap]

Originally posted by Fastracehorse@DRF
A speed figure is less work and just as effective as pace figures, ( IMO more effective ).

I cringe at all that work.

:cool:

fffastt

I use speed and pace figures and do fine using HTR's numbers. But I think I use them in a creative way that others do not.
Still don't know who has better figs yet, Cramer at HDW or Beyer.
I've done fine with the Beyers as one measure of my handicapping so I've got to do a study on Beyers-Cramer --- Except I don't want to buy the DRF if I'm already subscribing to HDW.

Besides my study would be different than others. I wouldn't
take best of last 3 or best last race and best 2 out of 3. That's irrelevant to me.
I would take 3rd best in the last 90 days and subtract 4 pts as a
race "par." Horses that can meet that par in their last 3 races within 6 months at today's surface (using only Fast/firm or good) and within 1 furlong of today's distance qualify.

Works better for older horses of course since 2 and 3 yr olds
improve to new heights more often, but I use form cycle
analysis to predict which horses could improve today.

[Fastracehorse]

I'd put my adjusted fig against anyone's - I would beat Cramer in ROI - it's an improvemnet of the Beyer.

If you use HTR/Cramer why don't you and I have a fun little contest over a few months picking 100 horses. I always luv showing the efficacy of a speed fig vs. pace figs/etc..

I know that sounds awfully fulllllllllll butttttttttttttt, I believe that much in the adjd fig.

fffastt

[delayjf]

fffast,

Sorry, but I'm not in a position to devote much time to handicapping right now and I don't currently have access to Cramer figures. I'm sure your adjusted beyer speed figure are good, but if you are adjusting your speed figures then I would contend that they are no longer speed figures but are now per've 'formance ratings (not that there is anything wrong with that)
While were on the subject, I've heard you say many times that the pace figure is a part of the Beyer figure, I would disagree. If I tell you that horse A ran beyers of 100,95,98, without looking at the racing form you could not tell me what the pace figures for those races are.

Understand I'm not doubting your methods, fact is nobody knows your methods because you won't share them, but that fine too. I'm not one to begrude another mans success.

My interest with GR1 was to see if the patterns outlined in Blinker Off could be discerned from the Cramer pace figures.
Perhaps GR1 will take you on.

Andicap,
I suspect that when the Cramer and the Beyers disagree, The Cramer numbers would prove superior. Cramer adds a lot of attention to detail to his varients that I suspect the Beyer guys lack. There are some who post here that worked for Cramer perhaps he could expand.

Best of luck

[cj]

I'm with you Fast, I'll put mine up against anyones. But only in my hands. The reason is I make them, I know exactly what goes into them, and what the strengths and weaknesses are. I don't think anyone can say that about publically available figures.

[Fastracehorse]

<If I tell you that horse A ran beyers of 100,95,98, without looking at the racing form you could not tell me what the pace figures for those races are.

+++++++++++++++++++++++++

That's kind of my point - If I believe that a speed fig has pace inherent in it's #, and it is successful, then why do I need pace #'s.

I was the guy that dissected every fraction too, not anymore - thank the horse gods.

fffastt

[Fastracehorse]

<But only in my hands. The reason is I make them, I know exactly what goes into them, and what the strengths and weaknesses are.

++++++++++++++++++++++++++++++++++++

To know your fig's weakness is important. It validates that you understand what it is doing and how you can travel the path of new discovery - further honing your skills, and your ROI.

fffastt

[Tom]

Horse wires the field and earns Beyer of 98.
Horse B closes from 8 lengths back at the half and is nosed out by A.
Both earn 98's.
When they come back in different races, don't each of those 98's have a different element of pace built in, unless you are using pace of race (which is the way I go).

BTW...agree with you and CJ-when you make the figs, you understand them and know them. Big plus.
But, as to HTR/Cramer, no way I can do that many tracks, so they are my pick for volume betting.

[delayjf]

CJ AND FFAST,

Perhaps a better contest would be between the two of you!
Maybe other figure makers like Game Theory want some of the action.

I have no doubt that your figures are better as they are home made. One day I might again decide to do mine by hand. Right now I'm going to take my que from other like Ken Massa, Dick Schmidt who have a have a lot of experience calculating figures / varients etc and have concluded the the added accuracy of home grown figures are not worth the added effort. They have been impressed with the attention to detail Cramer gives to his numbers. For now, I'll take their word for it.

Ken Massa (HTR) told me that given all the inaccuracies in the PP's that he feels that 35% (long term) is about the cap on a single horse win % using any single speed / pace number.

I am curious how many tracks do you play? I've made figures for one track at a time, but never multiple tracks.

And speaking of speed figures, CJ, were you ever able to determine if my timeform to Beyer Scale mentioned in any earlier post had any merit? Also, a while back you offered to e-mail your pace method to those who were interested, but for some reason we were never able to connect, I think you were in the process of PCSing back to the states. Any chance to take a look at it now?

Got to tell ya I was Jealous of you guys in Saratoga, I made my trip to Saratoga last year can't wait to get back maybe I'll hook up with you guys next year. How did Dave Schwartz hit that 100.00 horse?

[Fastracehorse]

Speed figs tell you how fffastt a horse ran.

How one determines if a certain style is beneficial ( early speed/ late runner ) is by 2 processes of handicapping: 1) Predicting the race-shape by lengths at call analysis and; 2) Bias of course.

If your late runner can run a 98 that's all you need to know - that he's a late runner. Simple. No fractions needed for that. No subscriptions either. Don't worry about it.

If their are several pace pressers and the track favours closers; well Tom..............it doesn't need to be difficult.

Right now Belmont dirt favours speed.

Santa Anita favours tactical speed.

A fffastt and easy way to predict biases is just go to Brisnet charts. If you do it for Belmont dirt races the last 2 days, you don't want to bet those 98 Beyer closers, unless you can predict a melt-down by lengths at call analysis.

Lengths at call analysis is just noting where horses are ( lengths behind leader ) at every call. Again, very simple. No prescriptions needed. Not much brain power either. If your horse is one length in front at the first call every race - again, chances are he'll be there again today.

:confused: :cool: :confused: :cool: :confused: :cool: :eek:

fffastt

[Fastracehorse]

Of my speed fig Delay.

Another advantage.

I hit longshot # 120 today ( $26 ) - hope to get # 121 in some time today too.

fffastt

[keilan]

Questions for ya -- of the 120 long shots hit this year 1)how many races have you posted? 2)how many horses have you picked to finish 1st which paid $10.00 or less.

[keilan]

Questions for ya -- of the 120 long shots hit this year 1)how many races have you posted? 2)how many horses have you picked to finish 1st which paid $10.00 or less.

[Tom]

Originally posted by delayjf

..... How did Dave Schwartz hit that 100.00 horse?

Magic 8 Ball (Chances are good...) :) :D

[Fastracehorse]

<Questions for ya -- of the 120 long shots hit this year 1)how many races have you posted? 2)how many horses have you picked to finish 1st which paid $10.00 or less.

++++++++++++++++++++++++++++++++++
++++++++++++++++++++++++++++++++++

1) I like to re-word question # 1 of yours. How many days do you think there are, where there isn't a 6-1 that wins?? And, how many of my second choices win at 6-1 or greater; further - how many of my top choices finish 2nd at long odds and pay great to place?? Most importantly: What is my ROI??

2) Over 2000 top picks that paid $ 13.90 or less. It's all on DRF if you want to check. Every longshot - every winner for that matter - has been archived.

fffastt

[Fastracehorse]

I have 2 posted top pick $100 winners this year.

One of them broke our daily contest record.

fffastt

[keilan]

If you don't want to answer the question -- just say you don't. It wouldn't offend me.

[Fastracehorse]

No offense taken - it's just that you didn't ask a question.

fffastt

[keilan]

Whatever -- you're too much work.

Forget it

[Fastracehorse]

What's up with that??

Now you owe me free drinks at the Gentleman's Club - on Gateway Boulevard.

fffastt