Two Preps

[TGJB]

Chris-- are you going to break down the sheets for the 2 prep horses? If you are, I\'ll hold my comments until then on the figures.

Jim-- we did a study going into last year\'s Derby on the question of new tops, because I was making the point about the horse almost always having to be fast enough already. From 97 to 03, 107 horses started, and only 8 ran a new top of more than one point (3 of which came in on 2 preps, even excluding the foreign horses-- VG, Lemon Drop Kid, and Proud Citizen). Last year, in that sea of slop, nobody ran a new top, so it\'s now 8 for 125, just over 6%. We don\'t have figures on pairups, might have by the end of the day. But Kevin posted yesterday that about 24% had run a pair or new top.

CH-- I don\'t know where to begin, especially since I\'ve already addressed most of the points, and you have once again completely ignored what I said.

So just briefly--

1-- Each Derby has 20 or so horses in it, and the figures of the horses are connected. While certainly there is subjectivity in assigning figures, there is no reason on earth to think we (or anyone else) would give horses with two preps better or worse figures than other horses. Which means while we can always get a figure wrong, any errors (ahem) or pattern differences should even out, and not favor any group that prepped any particular way.

2-- In a 20 horse field, the average chance of any horse winning is 5%. Over 50 examples you could easily come up with no winners or 6 winners, and still come out 5% in the long run (if you don\'t believe me, take a .300 hitter in baseball, and break his season down into groups of 50 at bats-- and that\'s at 30%). Even if we were to say winning was a better measuring stick than figures, I can\'t imagine how big a sample size you would need to test your thesis-- probably several hundred races. Any statistics majors out there?

Which is why I suggested using top 4 finishes (and why we print ITM along with win percentages in the trainer profiles). If something should on average happen 20-25% of the time, you can judge it with a much smaller sample size. Probably bigger than the one we have-- especially since training methods began to change-- but you should get a more reasonable result.

3-- Your comments about horses possibly moving up more with only two preps, but being slower because they have less racing, is not relevent. Since we use figures, we know how fast they are-- we know how they stack up against each other if they run their best, so the question is how they will run relative to their best. If they are \"behind the crop\" we\'ll see it in the figures. We\'re not going to be betting slow horses unless we have reason to believe they will improve, and slow horses ain\'t gonna win the Derby no matter how many preps they have.

If you respond to this, RESPOND to it, okay? Don\'t just say the same things over again.

TGJB,

1. I agree with you about your figures. I have simply tried to say that proving a point one way or the other using \"just figures\" could be open to interpretation.

ex. When a horse does not move forward after the Derby does that mean \"he fired a peak in the Derby\" or does it mean that \"not being prepared properly knocked him out\".  

I think there is no way to know. You and I might disagree on that point, but it might be key to determining if the horse was prepared properly.

2. I have agreed all along that the sample size is very small. In fact, that\'s the BASIS for me saying no one can be 100% sure of anything so the DEFAULT should be CAUTION. The idea of using in the money finishes has a lot of merit in many cases. I choose not to use it here because we are generally talking about horses that will NOT run terribly if 2 preps is not enough. We are talking about horses that will come up a little short - meaning they will often get a piece of the pie even if they are short. I have seen this phenomenon with layoffs on occasion. Doing the study would not satisfy me one way or the other based on my experience with other studiess.

Lacking a decent sample size is a major problem if the goal is proof, but several decades worth of horses is a start at evidence. We are not just limited to using the horses that TG can provide sheets for. We can look at decades of stats and look at sheets too. There is no law against using the stats available to form the beginnings of a view/evidence. The sample of sheet horses is also rather small.  

3. I am not thinking about it the same way as you.

Assume you have a horse running  5, 4, 2 (with the 2 being the Derby).

Assume the winner ran 6, 4, 2 1/2 -1 (with the -1) being the Derby.

I am saying that perhaps the pattern of the first horse would have been 5, 4, 2, -1 1/2 with one more prep. That would indicate that an extra prep could get a horse to run a faster peak on Derby day and would \"as a general rule\" indicate the proper way to bring a horse up to the race is with an extra prep. Without a scientific control we can\'t know how a horse would have run. We can only know how he did run. So I think we agree.  

My point here is the simple question \"is the pool of horses with only 2 preps at a statistical disadvantage because they generally aren\'t as developed\".  

My point here is \"not\" that the lack of development wouldn\'t be reflected in your figures. I know it would, but it doesn\'t tell us how to prep a horse.



Post Edited (03-30-05 15:44)

[P.Eckhart]

Not \'easily\', 7.7% for 0, 2.6% for 6.

I actually did make up a guesstimate of 2 (2 prep) runners per year at ave 17/1 over 57 years. Chances of no winners ocurring = 0.13%

[TGJB]

P. Eckhart (I knew a couple of Eckhart brothers who grew up here in Greenwich Village, one was really fast and played free safety on a team I coached)-- what are the breakdowns for 1,2,3,4 and 5? Even if you say there are only 6 possible outcomes, average for any one is 16%-- so 7.7 is not that low, or unlikely . Again, statistics are tricky.

In point of fact, Sunny\'s Halo did win off two preps (my first future book success at 40-1). And just since 82, when we started making figures, Victory Gallop ran better than the winner and lost because of ground loss.

[P.Eckhart]

For your example 0- 7.7%, 1- 20%, 2- 26%, 3- 22%, 4- 13.7%, 5- 6.6%, 6- 2.6%, 7- 0.8%, 8 or more 0.6%.

Yep I forgot there had been one winner.