I have been a satisfied customer for a while but there is something that has nagged at me.
Do you use different data bases for the trainer profile information on the left vs. the right hand side because even the number of events don\'t seem to fit.
Any insight?
No the databases don\'t differ but the basis for data in each differ. In the runs-based profile, left side, if the trainer and horse qualify, the run is tabulated. For instance, let\'s the horse is 2nd off a layoff, if the trainer had the horse prior to and up through this particular run, then the 2nd off the layoff criterion is achieved and the run is counted. Now, if the trainer claimed the horse in the preceding race, then the trainer DID NOT have the horse the entire time to satisfy the 2nd off the layoff criterion. (Be mindful, the horse itself, is 2nd off a layoff, but the focus and measure is trainer prowess, not what the horse does, individually.) Some of the criteria are race-specific and as such are counted every time for the trainer--the odds category for instance. But for other categories, 1st time lasix, is another, to satisfy the criteria, the trainer has to have the horse before and after.
The TG figure-based trainer profile, the right side, is predicated on more stringent criteria. The profiles are based on the \"effective top\' of each horse which is defined as the horse\'s best figure from its last 6 runs or within the last 12 months if it ran less than 6 times in that period. So for categories measuring how a trainer\'s horses do following a top, a pair of tops, etc., only horses that have run at least times for the trainer are used. Sequences where there was a trainer change in the last 3 starts are not used (except for the \"trainer change\" studies).
So for the figure-based profiles, we have to generate the patterns first and since the runs have to be for that trainer (for the most part), and trainer changes within the 6 runs, can invalidate lots of patterns, the quantity of observations that meet these criteria are much less.
I\'ve always wondered this. I know you guys do quarter points...as in a horse can run a 4.3..Do you really believe that these numbers are that exact? When handicapping...does a quarter point or a half point really matter?
The answer to the first part is we have to write down something. This way the leeway is small-- if we put down either 4 or 5 it would not be.
Actually, the fact that the TG figures are on a smaller scale than most others, I find to be an advantage. Beyer figures, for example are on a 3 times larger scale than TG. I find this has a tendency to exaggerate the importance of each Beyer point difference, which is really close to meaningless. A full point difference on the TG scale, on the other hand, is much more significant. Showing small differences as small superscript gives me a better picture of their lessor importance than full points.
In fact, when I used to make my own figures I used a a scale very similar to yours in which 1/5 sec. at 6 furlongs was equal to one point, which is very close to your scale in which a length at 5 furlongs equals one point.
Bob
We use 1/5 is a pt at 6f. You might have been using a slightly different beaten lengths chart-- at 5f a length is close to a point.
Correct. While I was using 1/5s rather than beaten lengths there could be some variation depending on how one values the relation between time and lengths. However, the scales are similar enough so that we would both value a fifth at 6F as 1 pt and 1 length at 5f as about 1 pt.