Maybe it's not relevant, as a manual handicapper, for me to be commenting. I've long known that as an experienced handicapper with the ability to discount or downgrade at least one well bet and one 5-1 or 6-1, ideally lower horse in a race, Bill Benter could have equaled or surpassed his self-described .24 ROI, as he mostly benefited from unsophisticated opponents and massive pools.
I received help in this area, though not for a model, but for quantifying a horse's trip concisely from a few generous posters...
http://www.paceadvantage.com/forum/s...ighlight=fuzzy
I've since greatly refined my approach, ending up with a 7 point Likert scale based on confidence and factor significance. One factor is a dead giveaway- being contrarian about public odds. If the favorite loses in today's race, say 67 % of the time, that's a lot of area for relevant information that will defeat him. But trips are far more subtle and significant than being blocked or checked.