Marc,

I got tired just sitting down to reply, since with every detail of my reply the points I will have to address in the future and number of replies I will have to make increase exponentially. I probably will have to cut off this discussion soon, because I don't have the time.

One of the reasons you appreciated my tone is that I avoided the "put every line in quotes and respond or comment" approach, which comes across as "every single thing you say is wrong and I will challenge every word", and plays as hostile, raising everyone's temperature. Sometimes that form is necessary, and sometimes hostility is called for, but not always.

A few points, before the 2/23 stuff specifically:

1- The test for the differences in methodologies come up in what you call the "most unusual of situations". When the track doesn't change speed and the sprints and routes don't split it's easy for everyone. My point was that by examining what happens when you follow Ragozin's methods (meaning assumptions) to their logical conclusion in the unusual cicumstances where the issue is actually relevant you can see the logical problems. Incidentally, the problem isn't with 4 1/2f races per se, it's with the ones at Kee that are run over a downhill course that older horses never set foot on.

2- There's a lot I could go after in your comments about drying out tracks, but I'll just say this-- I said ALLOWED for a reason. Ask Friedman, in public on his board, whether they have rules about how much their figure makers are ALLOWED to move the variant race to race, regardless of conditions. They don't measure those changes in track speed during the day, they pre-decide them by rule-- and by the way, that's how they came up with the Chilukki figure. It's the only way you could, since there was absolutely nothing else to go on. They worked back from the variant in the next race-- if it is X, then the preceding race has to be X plus or minus (roughly) 1 point. This is also why Friedman clammed up when I broke down his explanation for that figure-- he realized how that rule would play.

The way I know this, by the way, is that the guy sitting at the desk next to me worked in Ragozin's office for eleven years. As far as I know they did not have this rule back when I was there, but in those days Ragozin did all the "hard" figures himself, so he didn't need no stinkin' rules.

3- Yes, there was a more comprehensive explanation for FP's Derby figures-- "texture". That was the first time that scientific term was applied. In the real world, anyone who has made figures and looked at that years Ragozin Derby figures would see instantly that without FP pairing there was nothing else to justify the assigned figures, no reason to choose that particular variant for the race-- it wasn't tied to the other horses in the race, and there were no other races (this was a 1 1/4 race out of a chute) to tie it to. The figures for the DERBY were based on the disputed Wood figure, and the figures for lots of later races were based on the Derby.

4- "...when so many of the arguments are about drying out tracks and 4.5 furlong races, my skepticism remains high".

My arguments are about the METHODS used to make those figures, and logic (or lack of it). I did it with Storm Flag Flying's sept. race at Belmont, the Laurel day and others, because those show what happens when you try to tie together independent events.


On to the BG grass race:

Friedman did the race "right", given the data he was working with (the past histories of the horses). If you have looked at the figures I posted, you will see that I have also done it "right", for the same reason-- on both sheets the figures fit with the horses. But they can't both be RIGHT-- our figures run about 3 points faster, on average, and in this case they are about the same.

This is an example of what I meant by past bad figures creating further bad figures, whoever is wrong. In this case, it most likely is caused by the way each of us treat sprint/route relationships-- most of the time when they split you have to take off (give better numbers to) the routes. If you don't do this you will either use a variant that is an average of the sprint and route variants, giving both figures that are slightly off (sprints a little too fast, routes a little too slow), or tie it to one or the other, getting one right and the other wrong by a substantial amount.

As you can see from the figures we both gave the San Felipe and SA Derby, when you make the 3 point correction we have one the same, and one a little faster than Ragozin. The sprints, however, tell a different story-- in relative terms Ragozin has them much faster.

How do you guys know who is right? Over time you check the figures against results. But if you accept Ragozin's figures, it means the following: if BG had run as well as he did winning the grass sprint BY TWO LENGTHS, it would have been good enough to win both the San Felipe and Santa Anita Derby BY OVER SIX LENGTHS. Think about that.

Several posters (both you, Marc, and others on the Ragozin board) commented that the race looked fine on Ragozin. You are right-- it is internally consistent. It would be very interseting to get their (and your) reactions to the 2/22 Laurel races, evaluating them in the same terms. Maybe you can get him to post it.Only a cynic like me would think that Friedman's reasons for not posting one and posting the other would have to do with exactly the reasoning you guys applied in looking at the BG race...



TGJB