Quote:
Originally Posted by jay68802
I get the feeling I was not clear enough on my idea. What I want to do is build a data set of the horses the late money falls on. Then build or create a program that will pick the horses most likely to get hit by the late money.
Why not try to use what they are doing for improving my handicapping?
Can it be done, and how?
|
This idea is what I was referencing with the "thought experiment" comment. I don't know the answer, but intuitively, to me, if you're doing the above, you're in effect producing a replica of the models used by the large teams. If you're not predicting the price, and just trying to predict a yes/no answer to the question, "will late money fall on this horse", it might make the problem simpler, but I'm not sure and I lean towards doubting it. Perhaps it's easier to model, "is this a race that lends itself well to being handicapped by some model"?
With regards to your other question, doesn't being able to predict the final price assume that you DO know which horses they're going to bet?