Jerry,
I believe that the fact that a track can change speed is frequently obvious by just applying common sense to what’s happening to the figures. I know of at least one critic of your methodology, however, who claims that by doing so, you lose statistical significance.
As I know that you have a good knowledge of statistical methods, I was wondering if you have ever used or would consider Chi-square testing to show that there is good statistical evidence for splitting the variant or cutting a race loose on a given day. It is the method of choice for comparing 2 or more populations (races) to show if there is a difference in a variable (track speed). I think the results would be very interesting and would provide further justification that the track speed had indeed changed during the day.
Bob