|
I would toss the outliers. Your looking for the mean here or average for a particular classification of class not the exception. Ditto with other areas such as Distance or surface.
Class is more difficult to refine than say a Speed Rating. A SR applies to all whereas by class can be murky because it doesn't apply to all. It comes back to identical man made classifications have strong fields and weak fields. Even among G1 horses there are strong fields and weak fields, also from year to year among G1 can have a strong crop or a weak crop.
Look to capture the majority not the exceptions, this includes not only a runaway winner but also the winner against a very weak field and whose numbers don't meet the norm.
Much of this stuff particularity a model is not necessary to win and I don't keep any kind of model at all. If you do keep one do it on Factors such as V/DC, BLBL etc. Factors within the program will serve you much better. Even with Factors there will be outliers such as a horse that stretches out for the first time and so forth.
Sartin's later developments solved many of the problems through convergence of factors. There is a Follow Up on that but I don't recall the number, I believe it in the 70's and its on the front cover. Sartin even says a model isn't necessary.Back in the early years of Sartin, models etc. were necessary but not so with later developments.
Read my Beginners thread and I expound more on this in that thread and if your new it'll serve you well as its full of good info and current info. It's full of info that took many years to comprehend and I quote Sartin's word for word.
Enjoy,
Mitch44
|