Yeh tubby good thread that, some good iformation. I think IMO that finding the statistically imprtant factors is not the hardest part but more devising a systematic approach that weights each factor according to its merit, and finding the right balance that doesn't under/over compensate for a factor. The ratings i am trying to create are weight based and as such people suggest that 1kg = 1.5 lengths. But everything i investigate seems to be highly flawed. I will give you an example of what i mean. If you were looking at a factor such as weight increase. Obviously a weight increase disadavntages all horses to some degree. But a small filly is going to find 1kg a bigger burden than a large colt. This being the case i am finding it hard to adress these issues.

Cheers,
Matt