How much does hit rate deteriorate as odds rise?

[BJennet]

This is a very important question which I don't see addressed very often. When compiling statistics on handicapping methods or factors, handicappers will usually seek the average hit rate and odds of the relevant factors. One of the mistakes that is often seen is the assumption that the hit rate remains uniform throughout the full odds range. But anyone who has compiled such statistics in even a casual manner knows that the hit rate invariably drops as the odds rise. The key question is, do they rise out of the range of positive ROI.

Since we've been discussing NP extensively, we can use that as an example. If one looks at Bill Lyster's detailed 507 race survey of New Pace, the average hit rate for NP was roughly 75% (can't access this link from my present location). Yet, when one looks at the subset of longshot winners between 7-1 and 20-1, as I did, you see that the hit rate declined dramatically toward the rate of 2-2.5% per horse with the expected negative ROI. BTW, since this has been controversial, I'm not trying to single out NP - this happens with nearly every handicapping method and factor - positive ROI is the rare exception, not the typical result.

So, for those who are looking at their records, but don't usually consider this, one question you might want to ask is, not, 'what percentage of 4-1 shots did my handicapping hit', but rather, 'of my picks that went off at 4-1, did my hit rate yield positive ROI'.

Any feedback appreciated.

Cheers,

B Jennet