Jaguar
04-04-2002, 11:05 AM
Racecom's amazing Analog 5.0 owes its origins to Huey Mahl/Sartin methodology-but has been greatly enhanced by the addition of additional proprietary criteria.
Puzzling over the wonderful predictive power of this latest version of what started out as a truly terrible program(Analog 2.0)
I discern that Analog 5.0 is making models which include a clever "consistency" rating.
While Joseph and David Shepard would laugh at my bumbling efforts to understand what went into formulating this sophisticated and elegant product, I'll bet a guy like Dave Schwartz would understand Analog 5.0's elements right away.
Nonetheless, I daresay that this "consistency" rating bares some relation to the number of times a horse has finished 1st, 2nd, and 3rd- weighting these factors with a ratio which might be expressed as 11, 4, and 3-for example. Assigning 11 points for a win and 4 points for a place, and 3 points for a show.
That raw sum is then indexed by a class rating. I suspect that the net impact of this "consistency" rating is to shape the model of a given race type by assigning a value to the horse's performance within certain class categories.
Thus, when looking at Analog 5.0's output screen we do not see an obvious class rating(in traditional terms). Nevertheless, the horses listed have been measured for their performance at one or more class levels.
Joe Shepard once told me on the phone that in studying horse racing, they(Racecom) did not use any type of class analysis, since there was no evidence that class was a measurable or predictive tool. I suspect that is no longer the case.
Jaguar
Puzzling over the wonderful predictive power of this latest version of what started out as a truly terrible program(Analog 2.0)
I discern that Analog 5.0 is making models which include a clever "consistency" rating.
While Joseph and David Shepard would laugh at my bumbling efforts to understand what went into formulating this sophisticated and elegant product, I'll bet a guy like Dave Schwartz would understand Analog 5.0's elements right away.
Nonetheless, I daresay that this "consistency" rating bares some relation to the number of times a horse has finished 1st, 2nd, and 3rd- weighting these factors with a ratio which might be expressed as 11, 4, and 3-for example. Assigning 11 points for a win and 4 points for a place, and 3 points for a show.
That raw sum is then indexed by a class rating. I suspect that the net impact of this "consistency" rating is to shape the model of a given race type by assigning a value to the horse's performance within certain class categories.
Thus, when looking at Analog 5.0's output screen we do not see an obvious class rating(in traditional terms). Nevertheless, the horses listed have been measured for their performance at one or more class levels.
Joe Shepard once told me on the phone that in studying horse racing, they(Racecom) did not use any type of class analysis, since there was no evidence that class was a measurable or predictive tool. I suspect that is no longer the case.
Jaguar