All right, I just got in and saw this string. One of the problems with dealing with this the way it is presented here is that there are a lot of issues that range from related to semi-related to unrelated that have already come up in this discussion, and they have to be sorted through so the discussion can stay focused. As many of you know this is the stuff I'm always most eager to talk about, and I am more than willing to host a general discussion of this subject over the next few days. I'm going to start by making a few points now, and then I'm going to try to find some earlier posts that apply-- there are a lot, and I would be very happy if others can list some of them so the new guys can use the search engine on the site to look them up.

1-- As George (Mandown) pointed out, the only way to make figures is by using the previous figure histories of the horses. (By the way-- for those who don't know, let's establish George's credibility-- he co-founded the Racing Times with Steve Crist, among other things, and manages our data base. He knows what he's talking about, although it remains a mystery why the English can't speak English).

When you begin creating a data base, you create figures based on large population studies (par times, class levels etc.). Once you have a rough data base, you get rid of the pars (which aside from being non specific to the small group of horses you are measuring, usually are based only on the winners of the races), and start using ALL the horses in ALL the races to make your figures. This is true of all serious figure makers-- Beyer, TG, Ragozin. It's what is called the "projection" method, I think Crist came up with the name, but I could be wrong.

Also keep in mind that the figure relationship between horses in a race that has been run are fixed, by beaten lengths, ground loss, and weight-- we can't give one horse a figure we "want to" without firmly fixing all the other figures in the race in place.

2-- Anyone who has not already done so should listen carefully to my presentation from the DRF expo, which can be found on this site, and examine the scientific evidence I presented. You will come away understanding that for a lot of reasons, track "speed" does not remain constant throughout the day. So the question becomes how to evaluate the speed of the track as the day goes along-- making the ASSUMPTION the track stays the same is dead wrong, and results in creating and applying an AVERAGE variant for the day, which will give some horses figures that are too fast and others figures that are too slow. If you have one foot in a fire and the other in ice, on average you are nice and warm. Much more on this can be found in other posts.

For starters, take a look at my post of 6/28, "Friedman/Change of track speed", a lot of which applies here. LF had posted a response to a question by Class Handicapper, and this was my critique. Also, Jim-- point me to my exact quote about making the Derby figure, so I can see the context, and what I said.

More to follow.



TGJB