Okay, Del Mar opens this week, and with it the Thoro-Graph/Ragozin figure study. When this was brought up on the Ragozin board, Len Friedman said there should be 3 solid ground rules, and I agree:

1) The parameters of the study should be announced publicly in advance, meaning which tracks will be included, and for what time period.

2) The length of the study must be sufficient.

3) The study, once started, must be finished.

I agree with all those conditions. The tracks and time periods covered will be the Del Mar and Saratoga meets, which between them will have somewhere around 700 races.

Here's the way we will do it-- we will take the last 3 figures from both services run by every horse on the surface he will run on today (turf or dirt), no matter when they occured. We will throw out the worst one, and average the other two to create a rough power rating for each horse. If they have only run twice we'll average them, once, we'll use that one.

We will then adjust the ratings for the weight each horse carries in the race, and rank them in order of power ratings for each service. After the races are run, the service that has the winner ranked higher gets a point. At the end of the study, the side that has the most points wins. If the winner was a first time starter (or first grass or first time in this country, etc.), we go to the second horse, but no further-- it should be about who runs fast, not who runs okay.

As to how the mechanics will work-- George will put together a file and post it on this site daily in advance of racing with our ratings and rankings (and since we sell the sheets in advance and have the Red Board Room, anyone can check to see if we made a mistake or cheated). Tom (Alydar, Janis Joplin, etc.) will check what we post, and do the work to create comparable ratings and rankings for Ragozin, which he will post in advance on both sites and e-mail to me and Friedman. (What Tom is getting out of this is free sheets for both tracks from both services, which means I'm paying for a lot of Ragozin sheets-- ugh). Tom will also make the adjustments later for the overweights, and post a running score on both sites.

A few comments--

1-- I would like to keep things simple and focused, but I am open to any suggestions that Len has about how to do the study. There are issues here of both fairness and credibility.

2-- This is not a betting study-- we're doing it this way because patterns are very subjective, and different guys using the same data can (and do) have different opinions. But in the long run, fast horses will beat slower horses, and this study is designed to find who is more accurate in determining who the faster horses are. (Also, patterns are no more accurate than the figures they are based on, so if it becomes clear from the study that one set of figures is more accurate, conclusions can also be drawn about the accuracy of patterns).

3-- Because the data will be posted on an ongoing basis daily, anyone that wants to can do whatever other studies they want to (win %, ROI etc.). But the primary study is the one we are doing, because it creates a winner for one side in every race, making the sample size significant-- having to have the winner on top creates more randomness and more cause for argument (especially if someone bet the winner because they thought it "looked good", and it wasn't listed on top).

4-- Some of you stat guys might have an opinion about what would constitute a meaningful result, but in my opinion anything less than a 52/48 edge would not be significant enough for anyone to brag about. If someone ends up 50 points ahead, however...