Rags rationale vs. T-graph

[bobphilo]

Jerry,

I recall you wondering what rationale Rags uses for lumping all races together to come up with a single variant for the day. I remember from reading his book that he says that this gives him a larger sample from which to compute a variant. He even says that he will combine several days turf races to get a single variant. I can tell you from my training in stats and research methodology that large sample size is the most overrated factor amongst the untrained in research. Far more important is a representative sample, even if it's smaller. To lump together 2 different populations, such as horses running in races before and after a change in track condition due to wind, track maintenance, etc., gives a sample representative of neither population.
I used to use the Brisnet PP's, (partially because they were free with my Brisbet betting account) and the way they explain how they get their speed ratings is even more dogmatic than Rags. They also lump all races together (though they may split sprints and routes) and actually pride themselves that they just go with the computer generated variant with zero human judgment involved. Since there is no judgment involved, I suspect that they also use pars instead of projections. They also produce something they call Class ratings, which is similar to what you do when you break races loose, only more extreme. They basically rate a horse in the race relative to the other horses in the same race so it's like breaking every race loose. I found that the Class ratings are better than the Speed ratings, though they also suffer from poor projection. In effect you have the same outfit making both T-Graph and Rags style figures. Kinda shows how important it is to break races loose to get a more representative, although smaller sample. You might find it interesting to go to the Brisnet site and click on the library section to see this.

Bob  

[what about bob]

You can say what you want about Bris, but boxing their top three or four in every Breeders Cup race in ex/tris would have netted a person a very healthy ROI.

[jimbo66]

What About Bob,

How about you do that for a month or so and see if you get a healthy ROI.

There are days when boxing the 3 or 4 horses with the least amount of letters in their names would net profits.

My father wheeled the \"Italian\" horses in the Kentucky Derby and cashed big (Giacomo and Bandini).

[bobphilo]

LOL, Great title to your post.
I wasn\'t really trying to badmouth them as much as pointing out that their methodolgy for their speed ratings has some similarities to Rags and their class ratings to Jerry\'s. I still use their class ratings, esp when rating Euros (better than Timeforms) and give their speed ratings some consideration while understanding their limitations.

Bob

[NoCarolinaTony]

Now that I can relate with....

NC Tony