Last week there were a dozen or so posts on the Ragozin board about figure making, mostly taking shots at the methods used by Thoro-Graph and the other modern figure makers. I posted this just before I left for a few days, and there has been dead silence over there since. But I want to make sure everyone sees it, since the subject is sure to come up again.



There's been, for once, a lot of discussion aboput figure making on the Ragozin site, and now Friedman has actually chimed in with two posts. This would happen when I'm running around trying to get everything done before leaving for a few days in Florida, but he goes on the record so seldom that I have to take some time to comment. This is being done in a hurry, so I'm not quoting him in full-- if you go to the Ragozin site and the "Tale of Two Georges" (!!!) string, you can see his full comments, and know I'm not pulling him out of context. Any emphasis below is mine.

1-- "The variant maker then looks at all the horses who have run on the day and adjusts the variant level as indicated using the past histories of the horses (what I assume is meant in this discussion by the term "projection method" which I don't really like all that much as it implies that you'r looking at each horse's SHEET and projecting what you think the horse WOULD run off its history and then adjusting the level to accomplish what you EXPECTED to happen which is sort of a self-fulfilling approach that looks attractive-- as several others have pointed out-- particularly when the sampling is on the small side). This is not our approach to making variants".


Len-- Really? This is from page 34 of the book you co-authored with Ragozin:


"When I modify the variant, I look at how closely all the horses ran to what might have been EXPECTED, based on trends considering their recent form and their lifetime ability, and recognizing that young horses in good form may surpass their lifetime tops.
If a lot of horses are in line with what they SHOULD logically have run, setting the variant can be simple".


So much for saying that others base numbers on what they think the horses should run, and that you don't. ALL of us that make figures do, unless they are using ONLY pars, and even then you are basing your decisions on what you think a GROUP of horses should run.

The difference between our approaches is not that-- it's that you are willing to make some dogmatic assumptions, not supported by science, that connect unrelated events, using broad averages when doing so, while the serious modern figure makers (TG, Beyer, TimeForm) are not. Period. Beyond that, if we had a one race card to do, and no information other than the figure histories of the horses in that race, we would do the race exactly the same way-- choosing the most likely scenario based on what the horses have run, and figure to run.


2-- "The horses' histories do play the most important part in making the variant (although in a much less mechanical way than the 'projection method' implies to me), but weather and other factors that pertain to the track in question (and some that are general to all tracks) play an important part in the final variant decision".

Also from Len: "Within that framework,the horses (AND CONSIDERATIONS RELATING TO WEATHER AND THE PARTICULAR TRACK, ETC.) establish the level of the track variant for the day".

And along those lines, from page 35 of Ragozin's book, "Seriously underconsidered by Beyer and most other modern analysts is the LIKELY speed of the track today, based on weather-- especially precipitation-- and on the track superintendent's HABITUAL day-to-day changes in grooming the track".


Yikes.

a) First of all, does habitual mean always?

b) Not for nothing, but how did you determine the EFFECT of his "habitual" grooming, or weather, other than by looking at the horses and seeing how much faster or slower they ran than you would otherwise expect they SHOULD run?

And by the way, was it exactly the same every Friday, for example? Or did you work out and use an AVERAGE incorporating different effects on different days?

c) Regardless of how you got there, for God's sake, tell me how you would APPLY whatever you come up with. Would you simply do Thursday, and add 3 points to Friday because that's what usually happens, on average? Would you use an average variant you came up with for sloppy tracks? Or, WOULD YOU LOOK AT THE HORSES THAT RAN FRIDAY, AND BASE IT ON THEIR HISTORIES AND WHAT THEY FIGURED TO RUN? (Just as you did for Thursday).


3-- "There are situations where the track variant moves during the day (although much less than you and your cohort believe), but these situations do have 'reasons', weather related or otherwise".

a) Ahem. First of all, how do you know it changes, other than by looking to see whether the horses ran what they SHOULD have? And once you do that and see it changed, what difference does it make why? Shouldn't you just adjust the track speed (variant) for the AMOUNT it changed?

b) Yes, there are reasons, I talk about them in "Changing Track Speeds". But even aside from the subtle ones, there are obvious ones, like when the track is sealed in the middle of the card (like 7/27 Saratoga, when it was obvious to every figure maker but you what happened-- you gave all the horses in the two dirt races after they sealed it AWFUL numbers, because you tied those races to the ones over the harrowed track. Remember?). But of course, you can't know all the reasons if you make assumptions, and don't LOOK for reasons-- do you have trackmen record track maintenance information?


4-- "... most people have made their decisions one way or another, and I hope that those decisions haven't been made primarily by evaluating arguments but rather by seeing whose resulting numbers are the most useful in predicting future occurances".

Well, I can see why you wouldn't want anyone to listen to the arguments. But as for looking at the results-- with the NTRA contest finals coming up, I would like to point out again that our customers have won the last 3 of these. And one of the earlier winners is now using our data.


I unfortunately haven't got time to get into the "cutting loose" stuff now, because that is a very rich vein (think Jockey Club Gold Cup), but I have many times before, and I will again when I have time.







Edited 13 time(s). Last edit at 02/13/2006 12:01PM by nicely nicely.