Probability

[bobphilo]

Michael,

If the same lousy figure he earned running so poorly in the Sprint would have been good enough to win the Gold Cup, there is something obviuosly wrong with the figures. The Gold Cup figure was so low that even the figure earned in a crumby 8th place performance in the Sprint would have topped it. It\'s highly unlikey that this race could have been so slow relative to the other race, hence the figure is suspect.

Bob  

[Michael D.]

Bob,
One race has nothing to do with the other. 6F races will never validate or discredit the accuracy of 10f figures. Failing to address this point, you are simply giving an opinion that Ragozin got the JCGC too slow, with no evidence. While I appreciate your response, I was not looking for other poster\'s opinions on the JCGC figure.

[TGJB]

Michael-- Bob did define the point I was making. We all understand that horses can\'t all run all distances (although in this case Imperialism did run third {?} in the Derby at 1 1/4). But in theory identical speed figures are supposed to represent equivalent efforts at different distances. And if that\'s true, what Ragozin is saying is that Imp\'s BC effort (8th) was better than Borrego\'s Gold Cup effort-- if he ran to the same level at 1 1/4 he would have won the Gold Cup by 3 lengths, when you use the fractional points (1 1/2 points better than Borrego\'s Gold Cup).

[TGJB]

Bob-- the funny thing is that Davidrex\'s joke about Christianity was more to the point than sample size. As far as I can tell, they do it that way for the same reason that a lot of things in this business are done-- they\'ve always been done that way. Which is why I keep using the word dogma-- I\'m not kidding.

I have not looked at the Brisnet ratings, but the ones that are most similar conceptually in a lot of ways (and take it even further) are TimeForm. Those guys don\'t have fractions to work with, and many races in Europe have false paces, so final time is not relevant. Additionally, they have 3 day meets over screwy courses with hills, and very few races at comparable distances to work with. So they have to go by the horses-- the plus is that almost all races are on turf, and turf horses are very consistent.

[Michael D.]

\"And if that\'s true, what Ragozin is saying is that Imp\'s BC effort (8th) was better than Borrego\'s Gold Cup effort-- if he ran to the same level at 1 1/4 he would have won the Gold Cup by 3 lengths, when you use the fractional points (1 1/2 points better than Borrego\'s Gold Cup).\"

Jerry,
You say \"if he ran to the same level at 1 1/4\". Horses just don\'t duplicate 6f form at 10f very often. Why mix the two? Are these general performance figures or speed figures?

[TGJB]

Michael-- they are performance figures, since they include things like weight and ground. And what Ragozin is saying is that the same level of performance-- by Imperialism or someone else-- that was good enough to finish 8th in the BC Sprint, was good enough to win the Gold Cup by 3 lengths. And remember, Borrego KILLED that field.

But you don\'t have to look at Imperialism\'s 6f effort to get hold of this. The other 4 horses I mentioned ran at 1 1/4 miles both times-- do you think Suave ran well enough in the Classic to win the Gold Cup by 12 lengths?

Meanwhile, here\'s another one. You might recall that I did a pop quiz about Healthy Addiction before the BC-- pointing out that on Ragozin, she was the fastest filly (best top) going into the Distaff, based on her win (by a nose) over Fencelineneighbor in a restricted stake at Del Mar. He gave her 1 1/2 for that effort.

Same as he gave Pleasant Home for winning the Distaff.

Think Fencelineneighbor could have cleared Ashado and the others by nine on her best day?

[bobphilo]

Jerry,

Thanks for the reply. The explanation that Len overly emphasis the value of sample size in no why conflicts with the explantion that he is dogmatic. He just is very dogmatic in his beleif that a larger sample size is the main thing that matters - that\'s the rationale that he gives in his book. He, therefore, lumps in all the races to get a larger sample. You realize that a representaive sample is more important than just a large one so you\'re not afraid to break races loose to more accurately reflect the conditions present in a particular race.
I see the similarity of your figures with Timeform. They are, in effect, horse to horse comparisons of horses running under the same conditions, i.e. the same race. IMO, This minimizes the effect of weird final times and false paces becuse the main criteria is not the absolute final time or fractions, but instead the relation of the horses times to each other. They are a reflection of a horse\'s performance relative to other horses running under the same conditions. Unless I\'m mistaken, this is a lot like what you do.

Bob


 

[TGJB]

Bob-- Right.

This \"sample size\" question comes up a lot in different forms. Getting a large sampling is the basis for using pars (all 25k claimers), and the whole idea of the \"projection method\" is to get a representative sampling (the horses in THIS 25 claimer). One thing that people who haven\'t made figures generally don\'t realize is that while you usually only use the winners when using pars, you use all the figures of all the horses in a race (to various degrees) when using the projection method, which expands your sample size.

When you start a speed (or performance) figure data base, you have to go with large population averages (pars)-- you don\'t have figures for individual horses to work with. Once you have some broad data to work with, you begin to refine it in the specific (regression analysis, I believe-- Tony?) by using the histories of the horses that run in each race you are looking at.

In essence, Ragozin is making a mistake of logic (which doesn\'t enter into things if it is a matter of faith/dogma). He is using aspects of the broad average (par) approach in specific figure situations-- lumping things together to get quantity, not identical quality.

[bobphilo]

Michael,

I think the problem is you think that Jerry is saying the a horse who runs a given figure in a sprint is always capable of running the equivalent figure in a route. Of course this is, not true and would not be a valid way to compare figures. However, that\'s not what Jerry\'s point is based on. It\'s based on the assumption that the top level sprinters will run  approximately equivalent figures as top routers at the same level. If a methodology gives an 8th place sprinter a better rating than the top finishing router, than the figures are not showing proper equivalence- which is an important purpose of figures.
Let me put it another way. In baseball, an outfielder needs different skills than a 1st baseman but it is still possible to say one is a better ball player than the other. Let\'s say someone has come up with a way of rating players and, according to his figures, a 1st baseman who is in the bottom 10% of the league at his position is rated above the best outfielder in the league. Woudn\'t you think there was something wrong with how this guy was coming up with his figures?
Same principle here - a method that gives an 8th place finisher in a Grade 1 sprint a higher perfromance rating than the winner of a Grade 1 route does not  seem very accurate. It\'s either greatly overrating the sprinters or greatly underratting the routers. In this case the additional evidence from other horses in the race indicates that it was the route (the JCGP) figure that was to low.
Hope this clears it up.

Bob

[Michael D.]

Bob wrote: \"It\'s based on the assumption that the top level sprinters will run approximately equivalent figures as top routers at the same level.\" IMO, a horrible assumption (can of worms there).

I\'m out the door. Good discussion, I\'ll get back...............

I think the issue isn\'t whether a sprint figure should be \"theoretically\" equal to a route figure. IMO, it definitely should be even though they are different horses with different talents.

The questions are whether final time alone tells you how well the horses ran, who measured the final time better, did one of the figure makers incorporate something into his figure that the other didn\'t (which could account for the difference), do you want the figure maker to incorporate others things into the figure besides final time, ground loss and weight.  

[TGJB]

CH-- as you know, I deleted your previous post. I\'m letting this one stand as an abstract statement. You take one step down the path of making this another self-indulgent discussion of the things you always try to monopolize this board with, and you are gone. Try me.

I am trying my best to have a conversation that deals with \"ALL\" the issues without getting the post deleted or pissing you off, but that is very difficult in an environment where the first causes the other two. The truth is it is impossible to discuss the variant and figure issues here unless it is limited to the parameters that keep everyone from actually understanding ALL the issues.

[Caradoc]

I was going to try to bring a bit of diversity to the board today by raising the subject of the French riots, the causes and solutions.  However, I envisioned the following post in reply, and it just took the wind from my sails:

Most of the posters here haven’t studied riots and don’t understand the arguments and issues in coming to a proper analysis of their causes.  It is too tedious and time-consuming to explain to those who haven’t.  Whenever I try it, it isn\'t received very well anywhere.  When I post I usually anger some people and confuse the rest.  My points would get deleted or attacked anyway, and no one would appreciate it.

Oy vey . . .

[TGJB]

CH-- Yeah, that\'s the problem. Like I\'ve said before, you\'re a martyr.