"the right thing to do would be to change the pars-- you are supposed to end up with the winning figures for different distances coming out the same. Not for each individual class, and not necessarily with a specific small group of horses, but overall."

You are saying that the time chart would have to be changed if changed relationships could be demonstrated across the entire breed, correct?

That makes sense to me, but it might be more difficult to prove that than to just study specific groups of very versatile horses.

The overall quality of sprinters vs. routers could be different because of different financial incentives and other factors. That might confuse what the relationships should be for versatile horses.