[Dorianmode]

Originally Posted by rmath

Did you ever get a chance to run that comparison on DCL & LPR that you said you would?.

I just looked at that, and was working on an explanation for what I see. 96.2% of my top 3 horses HAVE LPR #1, and DCL #1 in the SAME horse, but with only 3 horses left to look at in RDSS, I guess that's not too surprising). For sprint races, one of either the top ranked RDSS horse OR the #1 LPR / DCL horse, in my final 3 contenders, win about 88.5 % of the time. On reviewing the races "missed", 4.3 % were marked as "mistake made in contender selection", (1578 races total in sample). So I have to go back through them, (the "lost" races), to see where the DCL/LPR #'s would sort out, had the contender selection been done correctly, before I can give you an accurate answer.

Right now, I'm spending my extra time keeping track of the "quality", (accuracy) of the morning lines at various tracks, to see if there is a significant difference, so it might be a while before I get to going through the bad contender selection races. By far, the most accurate single objective, (non-personal) contender selector so far, is "final odds" rank, but I'm trying to narrow it down further by crossing it with TE ranking of the final 5, 4, and 3, and eventually will do a sort by track. When the sample of the ML races is big enough, I will report the results. Also, am slowing working backwards, to add a few other possible elimination factors, and BY FAR, the most useful tool for eliminating the last horse from the 3 final contenders, is MY OWN pace handicapping, ("eyeballing") intuition/judgement. Was also surprised today that in my last 5 contenders, dropping the lowest APV horse seems to work, in getting the 5 down to 4, (I was NOT expecting that), but have a WAY too small a sample for that to be reliable, at this point. If you have any factors/ideas about anything you want added to the ML study/data, let me know.