[QUOTE=Bill Lyster;70178]
Quote:
Originally Posted by BJennet
Hi Ted,
Thanks for this detailed summary. Just wanted to add a couple of comments for newer clients based on my own research on TE. From a sample of nearly 1200 races the most useful finding for me (although maybe it shouldn't have been), was that TE rank is a significant predictor of win probability - each horse wins at a rate almost exactly 60% less than that of the TE rank above them. Interestingly Dave Schwartz came up with a nearly identical win rate for his prime factor, so possibly these ratios reflect something innate about the nature of how horses run."
Sorry it took me so long to read this, but I am having trouble with picturing the math you are referring to in the bolded section above. If you get a chance would you show an example of what you are saying. IS it, for example, that if the #1 TE wins at 60% that the #2 wins at 36% and the #3 TE wins at 21.6%?
thanks,
Bill
|
Hi Bill,
Sorry to have been so unclear. A couple of years ago, I did a study of TE, using a completely mechanical (at least as completely as I could manage) choice of pace lines: standard Sartin best TE of last three at comparable distance and surface (using judgement on timeframe). #1 TE produced 37% winners, #2 - 25%, #3- 15% , #4-10%. One important caveat: this sample eliminated horses displaying what I think of as 'false postive TE' - 'herd' horses that often hit the board but either never or rarely win, and NL types with the same win-averse behavior. Since I think these are easy-elimination types for most knowledgeable handicappers, I don't believe it should affect the significance of the results much, but I think it is worth mentioning for newer handicappers or new Sartin clients who might be reading this.
As I mentioned these are very similar to Dave Schwartz' results on his own prime figure, and I would guess, other overall measures of racing ability. Although my sample, at 1200, is on the low side, the variance of these hit rates was so low, regularly repeating at 50-race intervals, that I felt there was no reason to continue. One illustration of this phenomenon is a small Pick-6 I hit at DMR last year using only TE - three were #1 TE, two were #2, and one was #3. (Worth adding here that TE has added power on the NY and CA circuits).
I began this study after noticing that #1 TE seemed to be winning much more than with any previous iterations of the Sartin software. And compared with the win-rates listed in even the latest issues of the Follow-Up, my study seems to bear this out. As a result I began going over a sample of lost races, and found I had overlooked pacelines of winners with better TE in favor or pacelines horses that 'looked' better. As I began to use this mechanical method, my hit rate improved.
If you're familiar with Michael Lewis' 'Moneyball', in which he discusses the Oakland A's application of statistical techniques originally developed by people like Bill James to uncover previously hidden attributes which contribute to winning baseball games, this could be said to be somewhat similar. As Lewis mentions, even people considered 'experts' at evaluating talent - coaches, managers, scouts - were and are deceived by aspiring players who 'look' good in every way, but are less productive than seemingly
less athletic-looking types like Kirby Puckett. I believe something similar happens when handicappers look at past-performance lines - some seemingly bad looking pace lines and records contain some positive recent TE. This is what Doc was talking about when he said, 'Don't be deceived by the visual'.
One way of looking at this is to say that these are the frequencies at which the recent best performances in the field will repeat themselves.
Cheers,
B Jennet