I have to say, there is a decent discussion of pace, pars and figures on the Ragozin board, not that I agree with all of it. This is, to my memory, the first serious conversation about figure making to take place there since I started reading that board in 97.
Waaay back at the top, the original poster made a point, and asked a question-- yes, we had High Limit much faster than the other two figure makers, which is why we made the La. Derby ROTW. We had seen Beyer\'s figures and knew he had it off some (by about 3 of our points, 10 of his), and that Ragozin has that part of the country too slow in relation to others, so it seemed like a good way to make a point-- we had no idea he had HL as slow as the poster says he did, however (10\'s??).
Anyway, in answer to the question, yes, we had Afleet Alex running much faster at Delaware than Ragozin and Beyer did as well. His first two were 8 1/4, 2 1/2, and he came into the Sandford at Saratoga as substantially the fastest horse in the race.
There are two basic ways you can get a circuit (or several) out of whack. The first is to use pars, as, according to Friedman in a post from about 2000, Ragozin did until at least that time. At the Expo last year Len said they no longer did, which may or may not have been because of the dialogue I had with him about it in 2000-- it\'s also possible they were not using pars before either, and he didn\'t know it (he is a good handicapper but does not know a lot about figure making), or still are and he didn\'t say so at the Expo because he knew I would jump all over it. A related circumstance may be in play with Beyer-- he has implied they do not use pars, but that does not mean that some of the people who do circuits for him don\'t, and if they are he may or may not know it.
Which brings us to the other way circuits can get out of whack-- by having different people do them, who have different tendencies. We had a problem with this a while back ourselves-- the guy who was doing some of the smaller satellite tracks (River Downs, CT etc.) was on a different page than the other three of us who make figures (I do about the 8-10 biggest tracks, Paul and Greg do the next biggest). It became obvious when some horses left the circuits he was doing and \"jumped up\", so we went back and reviewed an awful lot of days, redid them, and fixed the problem. This is something we watch all the time, although it\'s tougher to catch when horses stay on one circuit.
Anyway-- as I have said quite a few times, Andy has had some circuits out of whack over the last few years (although to not nearly the degree Ragozin has). But Andy is extremely pragmatic, where Len is totally dogmatic, and from what I\'ve seen over the last couple of months Beyer may be adjusting, both currently and retroactively, as we did with those minor circuits. We\'ll see.
TGJB,
I\'ll repeat this again for everyone\'s benefit.
I have been told that Beyer does projection figures for all the major circuits. PARs are not used at all for any of the majors. I assume that means NY, CA, FL, KY and many of the others.
However, PARs \"are used\" for some of the minor tracks. Unfortunately I do not have a detailed list of which tracks fall into which category.
I am having dinner with a person that might know the answer to that question Saturday and I\'ll be happy to pass it along if he tells me. I am unsure how they do Delaware.
I do agree with what you saying in a general sense.
The thing that has me puzzled is that I also have been told that Beyer has a database of figures and software that can be used to monitor the figures of horses as they move from circuit to circuit. If anything is outside the specified parameters, it should get flagged for manual analysis. So these sorts of errors should get corrected fairly quickly if they are more than a length or so off.
Post Edited (03-16-05 17:19)
CH-- There are a lot of assumptions built into this kind of discussion-- that someone knows what they are talking about, that they are telling the truth, that what they are doing actually accomplishes what they intend (and claim) it does. This is a competitive business, and we are all going for it. As far as I know I\'m the only one who allows people to ask him questions about figure making, though the columns Beyer and Cardullo have written about figure making have helped de-mystify the process. The Kremlin down the block has done everything in their power to mystify the process-- to the degree they answer questions, the answer is basically \"Because we say so\".
What you have said basically fits with what I thought was going on, not that it matters. Don\'t know whether the program they are using is the reason, but from a cursory look they appear to have made some strides in fixing things recently.
Southern California and Delaware are good examples of why pars can get you in a lot of trouble, by the way. California has an inflated claiming structure and is geographically isolated-- the result is short fields of horses who would not be competitive at those levels elsewhere, and no source of horses to come in and take advantage. Gill had it exactly backward when he claimed horses there to take elsewhere (and that weakened the divisions even further). Delaware has a deflated claiming structure, lots of feeder tracks easily accessible, and drug barons who not only make horses run faster, but can run them aggressively knowing no one will claim them, further forcing other horses down to be competitive. If you really are using the projection method, none of this is a problem-- but if you use pars, EVEN AS A CHECK ON YOUR WORK, you will get in trouble.
And that is what I suspect Beyer has been doing, because of where (geographically) they have gone astray (as I said, they appear to be rectifying the situation now). That\'s where Ragozin used to get into trouble when I was there-- claimers from deflated areas like New England used to come in and blow them away at higher prices in New York and New Jersey. It\'s also where we got in trouble when we started 20 years ago-- and it\'s why we completely abandoned all par based figures many years ago. Pars are pretty much the only way to go when you start from scratch-- but you have to get rid of them as fast as possible.
TGJB,
My source about the Beyers is a mutual friend at the DRF. ;-)
I don\'t know how much about the process he knows but I am 100% certain that what he tells me is correct to the best of his knowledge.
PARs are almost certainly not being used in Southern CA. It\'s too important a track for any gambler to be using anything other than projection figures. I can\'t imagine that Beyer or anyone associated with him would ever bet a nickel based on PAR figures from CA.
Delaware is another matter. I doubt any of them bet Delaware too often and not that many horses from Delaware ship into NY, CA, NJ, FL, KY, MD etc....
Post Edited (03-16-05 20:32)
CH-- I agree that the guy in question is solid as a rock, but that doesn\'t mean he knows, even if he is told.
For example-- let\'s say you did a period of time day to day at a track by projection (meaning, using the figure histories of the horses who ran on those days as opposed to pars). But after that you look at the whole period and see you are running a couple of points fast or slow compared to \"par\" for the same period (meaning, on average you had the claimers etc. running a couple of points off what the averages would be). So you add or subtract from all the figures by a couple of points, under the assumption that helps your figures.
Did you use the projection method, or did you use pars?
Based on what Friedman said, this is basically what Ragozin did-- it\'s not that they used pars day to day, it\'s that they used them to keep their figures from going \"off track\". It\'s done because you don\'t have confidence in those making figures, and as I pointed out, it\'s a big mistake. I will guarantee you Ragozin himself didn\'t need no stinkin\' pars.
Anyway, that\'s what I suspect happened with Beyer.
TGJB,
>Did you use the projection method, or did you use pars?<
I\'d say you are doing projection figures and using PARs in an attempt to remove any biases that might creep into the process.
I agree that that can cause problems.
However, I think that not doing that can allow biases to run indefinitely. You could wind up with slowly shrinking or expanding figures that have nothing to do with reality even if they are all in sync.
I have yet to come up with a perfect solution.
As long as all the circuits are in sync, I think the scale itself doesn\'t matter much because any biases that exist take such a long time to screw things up the old figures are obsolete.
You definitely have to have some process in place that checks figures as horses move form circuit to circuit. That\'s the key.
CH-- of course I understand WHY someone thinks they should use the pars to keep the figures in line, but as I\'ve pointed out, if you do it you guarantee two problems-- your figures will not hold up circuit vs. circuit, and they will be useless for comparing horses from different generations, since you are starting with the assumption that the generations as a whole are equal.
Yes, you have to make sure the circuits match up. But you can\'t both do that and use pars, unless you are using flexible pars-- in which case they are not pars. Yes there is judgement involved, yes some figure makers have biases (Ragozin will not give out big numbers on a wet track, period, for absolutely no good reason, and has some other truly crazy ideas that I have discussed here at great length), but ultimately a figure is either right or it is not, in terms of the things it is supposed to measure. The trick is to get them right, and pars keep you from doing that-- as Friedman found out, and as I suspect Beyer is finding out.
Jerry,
I\'ve become very interested in Figure Development lately. Any good Books et al that you would recommend that would provide a good foundation? Have you ever written one?
Nc Tony
If you want to go historical, the book I have talked about here (see \"History Lesson\" in the archives) is great-- \"Consistent Handicapping Profits\", by Donaldson, first published in 1933. You can probably find a used copy on the web. And yes, it does have a \"Parallel Time Chart\" ( early speed chart), and a good explanation of how it works. It ain\'t perfect, but it ain\'t bad.
Other than that, Beyer\'s books and Ragozin\'s are good, up to a point. When dealing with Len\'s you have to look carefully at the assumptions, and completely disregard his account of the history-- to put it diplomatically, his memory is somewhat self--serving, and he takes a lot of liberties with the facts, in many cases by omission. But his account of how to set up a data base is good, up to the point where he talks about the assumptions and judgments going forward day to day and more importantly, race by race.
And if I was only doing figures for two tracks, I would consider writing a book too.
\"they will be useless for comparing horses from different generations\"
To be honest JB, I seriously doubt many real bettors care about this. I know that I don\'t. From a historical perspective, I\'d rather compare them on a \"par\" kind of scale anyway.
I think Babe Ruth\'s 60 homers are a much bigger accomplishment than Bonds 70+ when you look at what the rest of Ruth\'s competition did that year. With horses, I don\'t care if Ghostzapper ran a -6 while the other horses ran -4s, and Secretariat ran 2s. If the other horses of his era were running 8s, he is a better horse than Ghostzapper to me. What I want to know is how much better is a horse than the rest of its generation.
TGJB,
> But you can\'t both do that and use pars, unless you are using flexible pars-- in which case they are not pars. <
I thought that much was sort of understood - especially when it comes to minor tracks where the stock can change wildly from year to year.
When I used to make PARs just to get a grip on the \"quality\" of certain horses, I was always adjusting them from year to year.
NY Breds for example changed a lot over time.
TGJB,
>Ragozin will not give out big numbers on a wet track, period, for absolutely no good reason<
I don\'t know what \"his\" reason is, but there \"is\" a reason to handle wet tracks carefully.
The margins between horses tend to be much larger on very muddy/sloppy tracks than on fast tracks.
If the average winning margin at a specific distance is 2 lengths on a fast track, it might average 3, 4, or more on some sloppy tracks. The same is true for the gaps between 2nd and 3rd etc...
If you give the winner a huge figure to compensate for the slow figures the other horses are earning because they were beaten by so many lengths, you could be overstating the winning performace (upward bias). In many cases it may be more accurate to give the winner his typical figure and assume the others didn\'t like the off track as much and that\'s why they were beaten so badly.
No matter which way you go, this is an obvious technical problem with beaten lengths and off tracks that I noticed about 20 years ago. I am always suspect of the figures earned on days when it\'s clear that all the margins between horses were much larger than usual. I know that most of them are don\'t reflect the ability of the horses even if I don\'t know in what direction.
Post Edited (03-17-05 14:35)
Exactly. Meaning, if you use flexible pars you use the horses themselves to determine how fast they are, then turn that into pars, which you then use to determine how fast the horses are. All that does is add an extra layer of chance you can screw up by artificially setting standards, since those standards are based on your judgments anyway.
Once you have a data base of figure histories, you are always better off using the horses that ran in that race or on that day to make figures than some average based on another group of horses.
On off tracks-- those who don\'t work with the level and amount of really accurate data we do tend to understimate the sophistication of the decisions we make. If you saw an actual day, and how tight the figures are through a day, and the degree to which our choices are prescribed by a limited number of options once we have broken down beaten lengths, wind, ground, and weight, you would know that you are oversimplifying what we do tremendously.
Twenty years ago the tracks contained a higher percentage of clay, and when it got wet it got sticky, and stayed that way for several days as it dried out. That is exactly why many tracks went to a higher sand content, which acts very much like a fast track when wet (up to a point). I believe that Ragozin developed the habit he has back in the old days, when many more horses wouldn\'t try on the slower, sticky wet tracks, and dogmatic that he is, simply didn\'t adjust with the change in the tracks. There is absolutely no reason to believe that a significantly higher percentage of horses will not run on a track simply because it has water on or in it, if it\'s other characteristics are the same-- if anything, the sand may be easier to run on (think wet sand at the beach), which is why times are faster.
TGJB,
I mostly use PARs as a matter of quality analysis. It has nothing to do with speed figures.
Since I believe no figures are perfect, I tend to seperate horses with similar figures on other things. One of those is quality of competition.
If two horses have been running -1s and one has been running against Grade 1 horses and the other Grade 3 horses, all else being equal I\'ll take the Grade 1 horse all day long.
When it\'s graded stakes horses, quality issues are very easy because I know all the horses extremely well and I remember the depth of the field etc...
When it\'s statebreds vs. claimers vs. limited allowance fields, vs. races with age restrictions, turf vs. dirt, maidens etc... the quality of field issues can often get a little cloudy. Since I don\'t rely just on the numbers, I often answer these crossover quality issues by looking at the recent PARs for the class in addition to the actual fields in question.
That\'s why I spend time on it.
Post Edited (03-17-05 16:44)
JB,
Thanks!!
NC Tony