I'll spare you the blistering reply, P57 - an interesting post. The big problem with times is pace. As you alluded to, horses don't run the fastest time they possibly could, but most time-based calculations seem to be based on the assumption that they do. This is most clearly demonstrated with times becoming more unreliable the longer the race distance, however the unreliabilty exists even in shorter races. All this before you even consider the rather dodgy (in my opinion) area of calculating track variants, and the difficulty of accurately comparing times run at different tracks.

The question of weight could be argued back and forth forever. It can be demonstrated using physics equations that the change in weight has an effect, independant of the weight of the horse, or of the net weight carried. And at least you have the certaintly that a kilo is always a kilo, not some arbitrarily assigned values that are unavoidable in time-based calcs.

Your post begs the question, and I'd be interested in your reply - on what basis do you determine class?