David: Let me give you a clear target at which to aim:

I believe that a good rider makes it a bit more likely that his horse will run its best number or win with an inferior number (save ground).

How do we determine who is a good rider and who is not? Winning percentage cheats riders who get bad mounts and overrates riders who get good mounts.

After allowances are made for popularity or lack of same, ROI, because it incorporates the crowd's assessment of horses' chances, gives us a decent idea of who is good and who is not. There are better ways, to be sure, and we can talk about them if you like, but they take a lot more time.

Please note that I'm only interested in what ROI tells us about competence. For the purposes of a particular bet, I couldn't care less whether a particular rider is overrated or underrated by the public, just as, for the purposes of a particular bet, I couldn't care less what the takeout is, and just as, for the purposes of a particular bet, I couldn't care less whether speed figures are overbet these days. Later on, there will be time for jeremiads about how there were more overlays before the public had speed figures, etc. For now, all I'm interested in is how a horse's chances of winning compare to his odds.

The parts (read: the reasons why the public bets as it does) don't matter when we have the whole (read: the tote odds, against which we will put our estimate of the horses chances of winning) and confidence that our reasons are superior to the public's reasons. This is why I don't like analogies between handicapping and the stock market: A horse race gives us a result that has nothing to do with public perceptions. In the stock market, on the other hand, something like "greater fool effect" can reward a purchase that was wildly overpriced by any standard that excludes an expectation of irrational exuberance. It is as if bettors both handicapped horses and propelled them. I would have more fondness for an analogy between the stock market and a yearling sale.

If Bailey tends to improve his horses one point over a bad rider, that is important information. If the crowd knows this and bets Bailey's horses accordingly, I will have fewer overlays to look forward to, but the information retains its value, albeit more as a prerequisite for success than as any sort of guarantee of success. Before I bet a race, I look at figures, pace, riders, trainers, etc., and make an odds line in my head. Later, I will compare this odds line to the tote odds and bet on horses that I think are overlays.

If Bailey figures to move a horse up by one point, but the crowd bets Bailey as if he will move the horse up by two points, the horse can still be an overlay and a great bet, depending on whether other factors overpower the one-point-versus-two-points discrepancy.

Any talk that a handicapping factor, whether it's riders, TG figures, or anything else, is unimportant because the crowd will take it into consideration and bet accordingly, strikes me as epic, monumental, awe-inspiring sanguinary BS. The percentages that we give to individual horses--reflecting their chances of winning, one hopes--must add up to 100. This is a zero-sum game. To refuse to account for a rider's ability to move up his horses because the crowd bets his mounts accordingly is to guarantee an error elsewhere. It is to double count, one count being the expectation that Bailey's talent will be negated by lower odds and another count being the actualization of the negation on the tote board. (Sorry about the rhyme; Bulworth mode; great movie; Halle Berry at her most ravishing.) In short, it is to botch the operation all to hell.