Sidebar

[DeltaLover]

Originally Posted by steveb

while essentially that may be true, there is still two things out there in the race......a horse AND a rider.
i have always been of the opinion the rider is the more important of the two.
and your chart could be influenced by one or two long priced winners?
not to mention it is very fluid, so it probably looks very different month to month.
or you could have riders that are profitable......depending on how you filter, but you have less than 1.
there is seven zillion ways to analyse it.
you could have a rider with negative expectation, but in the circumstances could be a great bet.
rider IS VERY very important, but it's just one factor.

personally i would never bother analysing a rider the way you have done. it does not work that way.
if all your factors as one give positive expectation then what does it matter if the rider in this case, is a losing proposition according to his/her past history.
you job is to predict the future not the past even if the past may help to an extent.

You are right in the sense that a flat ROI analysis like the one I present here leaves a lot of room of improvement; removing outliers and applying a moving window based on time are two of the ways to do so. More that this though your job is to predict the error in the crowd's estimate and not the most probable outcome of the race; this is the point I am trying to make here.