incremental odds analysis

[formula_2002]

I find it quite astonishing that so little discussion is devoted to meaningful, increment odds analysis.

Pace figures, speed figures, all sorts of data from Bris, Equibase, Daily Racing Form..and so on and so on, are discussed in micro-opic detail, but so little effort is devoted to the analysis of the odds.

Just look to publications by Benter, Fabricand, Quirin. They include the incremental odds analysis.

Perhaps we need a new forum, “INCREMENTAL ODDS ANALYSIS”.

Joe M
PS.This type of analysis has save me a considerable amount of money..My largest cost is the data

[ceejay]

What do you mean by "incremental odds analysis?" Things like ROI between 6- and 8-1 in a 10-horse field?

[kenwoodallpromos]

Ainslie had an odds analysis, and the new book "early Speed" has it- but you have to understand how to interpret the more detailed analysis to make good use of them. Even Klein's the emphasis (of course) is an early speed without regard to other subjective factors. and depending on what other factors the public uses in its' handicapping or betting, a set odds table could change over time or between tracks.
But I think ball park figures are very useful- maybe that is why they are not out there too much! We do not want everyone betting based on actual "fair odds" and and accurate figures and the tracks would not want that done either!! (rerence CJ's limited sales of his figures, which are the most accurate of all).
Look at the M/L for yesterday's GGF short-field day- Richard Somers still put routine 2-1, 5-2, 3-1 odds for the top 3 horses, even in several 5-horse fields!!LOL!!
The actual final odds yesterday compared to the M/L and compared to the order of finish are quite different.

[Tom]

Can you distinguish the difference in performance of a 3-1 shot compared to 3-1 shot that is a 75% overlay?
If not, who care - nobody is going to bet every 3-1 shot, so your results are not representative of the potential outcones.

[formula_2002]

Originally Posted by ceejay

What do you mean by "incremental odds analysis?" Things like ROI between 6- and 8-1 in a 10-horse field?

6-1 odds, as a % of pool is 14.285%
10-1 odds, as a % of pool is 9.09 %
In my thinking, that is too large a spread to work with.
A 6.1 to a 6.9 spread is 1.33 %. That may be more meaningful.

I would like to be able to compare the incremental odds results for any horse racing system to two things.

1. The public’s win % in the various odds range were.

2. Random number generator win % in the various odds range.

Then, using a simple formula for statistical significance, determine the significance and profit potential of the system.

[formula_2002]

Originally Posted by kenwoodallpromos

(rerence CJ's limited sales of his figures, which are the most accurate of all).

With much respect for CJ and his work, I ran an increment odds ananlysis on his performance and speed figures. At this point in the study, it would appear that the speed figure is superior to the performanc figure.
There were 11,000 plays in each study.
I sent CJ a copy.

I also did a similar analysis on Jeff Platt's JCAPPER program. Amongst other thing his programs also offers a single figure along with an odds line.
That combination produced the best results I have ever been able to generate..

[formula_2002]

Originally Posted by Tom

Can you distinguish the difference in performance of a 3-1 shot compared to 3-1 shot that is a 75% overlay?
If not, who care - nobody is going to bet every 3-1 shot, so your results are not representative of the potential outcones.

That's another thing that interest me.
I was just working on that very thing.
I was using 4-1 horses.
Almost 15,000 horses in the range >=4<5-1 produced a lost of (as expected) 19%.
If certain on these horses were overlayed by 75% in each of the 10 cent incremental ranges between 4 and 5-1, the profit would be 44%.

But that raises the question;"What is the maxium number of these horses that can be overlayed 75% without impacting the entire range which must lose 19%?

[twindouble]

Originally Posted by formula_2002

That's another thing that interest me.
I was just working on that very thing.
I was using 4-1 horses.
Almost 15,000 horses in the range >=4<5-1 produced a lost of (as expected) 19%.
If certain on these horses were overlayed by 75% in each of the 10 cent incremental ranges between 4 and 5-1, the profit would be 44%.

But that raises the question;"What is the maxium number of these horses that can be overlayed 75% without impacting the entire range which must lose 19%?

I say, pick the winner and take from there.

[cj]

Originally Posted by formula_2002

With much respect for CJ and his work, I ran an increment odds ananlysis on his performance and speed figures. At this point in the study, it would appear that the speed figure is superior to the performanc figure.
There were 11,000 plays in each study.
I sent CJ a copy.

I also did a similar analysis on Jeff Platt's JCAPPER program. Amongst other thing his programs also offers a single figure along with an odds line.
That combination produced the best results I have ever been able to generate..

Joe sent me the data as he said, but I have absolutely no idea how to interpet it, or even what went into it. All I know is that for every $1 I bet, I get back about $1.07 $1.08, before rebates. All top figures are not created equal. For example, I just made my biggest bet of the day on the winner of the 8th at Pimlico, a horse that was no better than third on my figures going by the last race:

64736494-1
32 1:29pm
22-Apr-06 22-Apr-06 Horses
Win Pimlico, Race #8
6.Pagan Moon
200.00 700.00

Yes, this is a redboard, but that is not the point of this post. Actually, in the interest of full disclosure, I had missed 15 races in a row before this one today.

Instead of incremental odds analysis, maybe incremental race analysis, breaking races up by sex, age, distance, surface, and which figures perform best in which category of races. Not only that, how about pressured races, or lone E races? Races with unknowns, such as first time starters or first time turfers? There is soooo much more to this game than what you are attempting to measure, though I do respect the effort.

[kenwoodallpromos]

I may have mentioned this prior, but I would like to see an analysis of win payoffs (odds) in ranges of = under $6, $6-$9.99, $10-$17.99, $18-$25.99, $26 and up (or similar brackets).
Variables would be the major race conditions of distance, purse or price level, field size, post position, and win recency.
The possiblilty of longer odds horses winning to certain race conditions is the most reliable blanket angle I have run across.

[Murph]

Ken and CJ, you are aware that you can do this type of anaylisis, based on Thorostats selections, using the past performance database, I hope. If not please check it out and let us know what you think. Does this tool help meet your needs to classify the results by ODDS?
Murph

[formula_2002]

Originally Posted by Murph

Ken and CJ, you are aware that you can do this type of anaylisis, based on Thorostats selections, using the past performance database, I hope. If not please check it out and let us know what you think. Does this tool help meet your needs to classify the results by ODDS?
Murph

Murph, could you post a sample of your incremental odds analysis? I'd like to see the standard deviation performance for the factors. The back up would also be helpful.

Thanks
Joe M

[rrbauer]

I guess that the incremental odds analysis might make for some good academic exercises. I'm wondering that given the amount of late money that shows up long past the "who should I bet" decision-time-window; how practical is having the knowledge?

Anybody?

[Murph]

Originally Posted by formula_2002

Murph, could you post a sample of your incremental odds analysis? I'd like to see the standard deviation performance for the factors. The back up would also be helpful.

Thanks
Joe M

Hello Joe,
I can post up somthing similar to what Ken is asking about later today. In the meantime our site is available to you for an e-mail address. We won't spam ya, promise. If you are busy or whatever, I'll post later with a bit of what I look at with a screenshot. Here is a thread with an example of the input form and sample output. This example isn't adjusted for odds, though you can see where the option is available.
http://www.paceadvantage.com/forum/s...908#post261908

First I have some chores to do this afternoon, let's see - KILL ivy, attack lilac bush and dead prune the roses - then mow the yard. I'll be back around dark!

BTW - I couldn't log into the WAR ROOM Fri afternoon between 4 and 4:30 or so before I gave up. I will try again tonite to see what's up with that and let PA know if I got a problem. Anyone else get this lately?
Murph

[kenwoodallpromos]

Thanks for the pointer to your selection odds- I guess I am just lazy, was looking for compiled redboard stats!

Murph- if someone wanted to pay you for the use of your power rating (and include a link to your site) to market along with other non- DRF/Equibase/Trackmaster ratings/figures with raceday program information, would you consider it?
I'm sure you know better than I who the "independents" are.
Is anyone already doing that?
To get back on topic, maybe statistic could be included with the race header as to % of winners for that particular type of race broken down by odds brackets? I do not believe thast would violate ANY Actual copyright laws and would promote alternative numbers maker such as yourself.
Or you and opther could do what NTRA did- form a non-profit company to market it!