Probability

[TGJB]

Some of this has been covered in various posts, but I want to get this clear and on the record, so here goes.

The point of making figures is to assign the ones that are most likely to be accurate, to be the most accurate representation of how the horses performed, given what it is you are supposed to be measuring. The only way we have to do that is by using the past figure histories of the horses that ran in the races we are assigning figures for, and that part is not in dispute between Ragozin and myself. The argument comes in deciding exactly which past figures to look at-- specifically, whether it is correct to ASSUME that exactly the same conditions exist for different races that you are assigning figures for. I have shown as a factual matter that Ragozin\'s dogmatic assertion that that is a good idea is wrong-- both with examples like the separate watering of the two Belmont grass courses (Len combined them for variant purposes as a matter of course), and with the scientific evidence of those who have studied the physical properties of actual racetracks (The Expo presentation in the archives section,and some of \"Are Racehorses Getting Faster\"). But the best evidence for those who do not make figures comes not with theoretical discussions, but with examples.

Briefly, for those that have not followed the discussion, if you did all the dirt races on Breeders\' Cup day assuming track speed stayed constant, the Distaff came up very slow. It looks like Ragozin-- who does NY himself-- may have taken off about a point from that race himself, but that also could be a function of slight differences in our speed chart, wind estimates, or wind formulas, (especially given the long straightaway). Regardless, he essentially did the race with the rest of the day, not cutting it loose like I did-- which I would have done even if I did not know that they had watered the track only once during the day, and the Distaff was the first dirt race after that.

I posted this before, but I want to get it all in one post, so I\'m going to go over it again. ON RAGOZIN\'S FIGURES, not mine, the field for the Distaff had run 74 times this year, and 33 times had run more than 3 points worse than what their top was at any given time (45%). That\'s a pretty decent sampling-- if you did a similar study for previous Distaffs you would probably find something in the same general ballpark. Well, the way Len did the race, he gave 12 of the 13 worse than 3 points off their tops (and 9 of those more than 6 points off their tops, by the way). There has been quite a bit of discussion on this site as to the right way to calculate the chances of all 12 running that bad, with the most generous being Jimbo\'s at one in 2 thousand.

Which doesn\'t seem like the most likely configuration, the one that is most likely to be accurate.

Now, one could take the position (as Friedman did at the Expo) that all permutations are possible over time. That may be, but you are still supposed to come up with the scenario that is MOST LIKELY, given all the available information (like watering). And since we know from the science that making assumptions about the track staying the same, even without having specific information, is a bad idea (that\'s what got Len into trouble with the 2 Belmont grass courses) there is no basis for using that assumption when it directly opposes the histories of the horses in the race or races in question. Conditions might be the same, and might not, but ASSUMING either is a mistake.

But lets say you accept Len\'s position, which is that this 2,000-1 shot came in. If so, it wouldn\'t happen too often, would it?

Well, here\'s the thing. Just from looking at Ragozin\'s BC sheets-- and not looking too hard at that-- there are two other obvious examples of the same kind of thing. One is the last 2 dirt races on 7/27 (opening day) at Saratoga, where they sealed the track in the middle of the card, and Friedman said that didn\'t matter, it was right to treat all the dirt races as being run over exactly the same track. Adieu and Folklore came out of those races:

Folklore (earliest first, and ommitting fractions)  21-14-13-18-10-5-9
Adieu                                               11-14-20-9-14-16

The 18 for Folklore is a WIN on 7/27 in the GII Adirondack. Adieu was a close fourth, and got a 20. And all the horses that ran in those 2 races (with the exception of the winner of the other one) will look like that. Which is why I predicted AT THE TIME that they would all \"jump forward\" next time on Ragozin, like Folklore and Adieu did (8 and 11 points respectively).

The other, of course, is the Gold Cup. Last 5 for the ones that I have sheets on, because they ran back in the BC, next to last being the Gold Cup:

Imperialism        3-8-1-13-4
Flower Alley       3-0-0-15-0
Suave              5-6-3- 9-1
Sun King           5-5-2-11-6
Borrego            8-5-2- 5-4

This is not the whole field, but the ones that I don\'t have were beaten even worse than the top 4, and got even worse figures. So every starter (INCLUDING Borrego) ran at least 3 points worse than his top, on Ragozin, in the Gold Cup. And the ones other than Borrego ran MUCH worse.

Mathematically, what is the chance of these examples happening with just  the small group running in the BC races?

That aside-- I\'ve done this before to make a point, but I\'m going to do it again to make sure it\'s clear:

If you believe those Ragozin numbers, you are saying that if Imperialism ran as \"well\" as he did in the BC (he finished 8th in the Sprint) one race earlier, he would have won the Gold Cup.

You are saying that if Flower Alley had run his BC race one start earlier, he would have beaten Borrego in the Gold Cup by 15 lengths (and the third finisher by 25 lengths).

Same thing for Suave-- he ran fifth in the Classic, and that effort would have won the Gold Cup by 12 lengths.

And of course, you are saying that Borrego ran better when finishing 10th, beaten 10 lengths, than in winning the Gold Cup by 4.

[bobphilo]

Jerry,

I know you\'ve been busy lately, but I was curious whether you agree with my interpretation that the difference in the philosophies of you and Len (in my post of 11/8 re: Rags rationale) lies in that fact that Len is mainly concerned with sample size while you are more concerned with a representative sample?
Also wondering if you got a chance to look at how the Brisnet class and speed  ratings seem to reflect the respective methodologies of you and Len.

Bob  

[JAKE]

I added a few numbers

Rags  .......................................           Thoro

Imperialism 3-8-1-13-4 .............          1, 2, 1, 2, 1
Flower Alley 3-0-0-15-0 ........ .       -1, -2, -2, 5, -2
Suave 5-6-3- 9-1 ................               1, 2, 0, 0, -1
Sun King 5-5-2-11-6  ..............            0, 0, -1, 2, 2
Borrego 8-5-2- 5-4  ...............             2, 0, 0, -3, 0
Perfect Drift 4-6-1-6-2  ..........        0, 0, 0, 0, -2
Sir Shackleton 9-7-5-3-5 ..........        2, 0, 0, 0, 0


My question is \"What is the probability that these horses not bounce?\"
According to your numbers, they don\'t bounce (hardly ever) which is not believable. Horses are not machines. They are athletes that have ups and downs as any athlete does.

The point of \"if Imperialism ran the 4 in the Gold Cup he would have won\" is nonsense. I guess you would also say if Silver Train ran his 0+ Sheet # for his win the the BC Sprint in the JCGC, he would have won by a mile. Those comparisons can\'t be made. It just doesn\'t make sense.

[Josephus]

Jake,

Flower Alley bounced from a -2 to a 5; that\'s a 7 pt. bounce.

Sun King bounced from a -1 to a 2; that\'s a 3 pt. bounce.

Borrego bounced from a -3 to a 0; that\'s a 3 pt. bounce.

Am I missing something, how does Silver Train\'s 0+ in a SPRINT translate to a \"win by a mile\" in the JCGC when the winner ran a -3?  Yes, horses aren\'t machines, but the good ones are pretty consistent---that\'s what makes them good and these are the best ones still racing.  But please don\'t change your opinion, it helps our prices.

[jimbo66]

Jake,

If you really want to make a good counterpoint to Jerry\'s posting, then point to a figure or figures that Tgraph gave to a horse that doesn\'t make sense. There are people here that will listen to your point and consider it.  This ain\'t the Ragozin board with a bunch of groupies all claiming to hit every race because of Len\'s great figures.... Making a general statement about too many pairups is a statement lots of people here have heard before.  

The examples Jerry gave are pretty poignant, even if nobody who uses Rags is listening.  Jake, do you really believe that Borrego paired up his JC Gold Cup race in the Breeders Cup?  Do you really believe 12 of 13 horses in the Distaff bounced 3 points or more?  If you do, that is your prerogative.  If the Rags users all believe that, then Len Friedman would do Jim Jones proud, he must have unbelievable personal charisma.  You guys are beyond giving \"blind faith\" to the figures.  Don\'t they have to make sense to you Jake?  I assume if you are paying $25 for a product you have a sophisticated enough level of understanding of the game to see that something makes no sense.  

[JAKE]

I do believe that the BC Classic was a much faster race than the JCGC.

How Thoro has Borrego reacting from the JCGC to the BC is a # that I disagree with.

How Thoro has Perfect Drift\'s race prior to the BC as a 0 on an off track is another # that I disagree with. Was that effort as good as his Pacific classic race? No.

The 0+ of Silver train was being compared to the 5\" of Borrego (Rag #) not the
-3 of Thoro.

Did Imperialism only bounce 1 point (from a 1 to a 2) from his 7 furlong sprint win off the layoff to his JCGC race? No.

I could go on....

I\'m not saying Rags is 100% correct, but neither is Thoro.

[brokerstip]

Jimbo,

That is a very fair and \'calm\' response.

Jerry made a very compelling argument. After reading all the facts over the years, I am very impressed with the insights Jerry regularly provides on this site. I am also quite impressed with his willingness to address specific issues such as the recent discussions on the Cal. figs.  
Any objective player of \'numbers\' whether they are Beyers, Rags, Fotias, Henry Kuck etc. etc. has to be impressed with Jerry\'s attention to detail, his constant search for the \'truth\' (all the work with watering, cushions, Porcelli etc.).

Some time back, I called the TG office with what I thought was a Sam Houston discrepancy. ( a turf race that looked like Rags Gold Cup--a bunch of horses that all had huge bounce numbers off the same race). They checked it and corrected it ...(it was no big deal). I liked the way they handled the question and the professionalism they exhibited.

As someone who has used both products and has a lot of respect for LF, JH and crew,  one would have to be blind not to see that Rags has the Gold Cup plain wrong.

Jimbo\'s suggestion that one refutes Jerry with specific examples is spot on.  

[gohorse10]

This is my first post. I have been betting horses for 40yrs. The first bet I made was at Monmouth I was 5yrs old 5 to win on Mr. Brick he won. I started using the sheets(Rags.) about 15yrs ago.I spent alot of time at Pha.& Garden state with all rags users. I drove from central NJ 70 miles because Norm took good care of his rags customers.I did use TG at times also. I thought rags had better numbers than TG when they were done by hand.I changed to TG when he went to the bound hard copy book.It was a better product & a better deal than rags. IMO this is when TG started to leave rags behind. The discrepancy in the numbers between TG & rags have been going on for yrs.I found this out at Saratoga because I used TG but still new a lot or rags user\'s wanted nothing to do with TG but I still got to look at rags sheets also. I would find big differences in some horse numbers 75% of the time TG would be a lot lower.On TG you would use the horse & rags he would be a toss.The TG horse would run to his number & would win or be in the number.If you had rags you wouldn\'t have the horse.I spent yrs trying to convince the rags people to switch to TG by showing them the discrepancy between the products.I have converted a few but the rags people are very loyal.So the point is TG has left rags in the dust. If rags has been making mistakes for yrs what does that say about his numbers today or tomorrow.For the TG user it is to are benefit if rags doesn\'t have it right.JB keep up the good work & thanks for a great product.

[Michael D.]

Jake,
the super in the classic paid over $12g. the favorite won, and one of the horses that took money came in 2nd. you say the TG had the Hawthorne GC too fast. well, two horses that came out of that race rounded out that super, one at 14-1, and one at 70-1 (longest shot on the board). do you really want that Haw figure slower? you are less likely to come up with super frolic and the $12g super if you make that race slow.

Michael,

\"do you really want that Haw figure slower?\"

I thought he was implying that the Pacific Classic should be faster and that Perfect Drift ran poorly in the slop at HAW.

Make the P.C. faster and....

PD\'s number in the BC makes more sense.

Super Frolic\'s overall sheet is more consistent.

Borrego looks better coming into the JCGC.  

Of course if you did that, someone like me might say that Super Frolic ran \"great\" that day given his trip. Wait a minute I already did that and no one agreed. :-)

[davidrex]

Kudos to c.h. and Go horse for their respective threads.

Problem I have is with all the \"handicappers\" on this board.

Yes t.g. has moved to another plateau...and yes its like pulling teeth to convince  a raggie (much less flipping one),but why is it so damn important to prove this point when as a raggie has already stated on this board \"if t.g. is so good,why do raggies send so much more money thru the windows?\"

Can\'t you see that winning this battle can only cost us all money down the road?...after all this isn\'t Christianity your fighting about.

PARTYpokerON!

[Michael D.]

\"If you believe those Ragozin numbers, you are saying that if Imperialism ran as \"well\" as he did in the BC (he finished 8th in the Sprint) one race earlier, he would have won the Gold Cup.\"

Jerry,
I am confused here. Could you elaborate. Thanks.

david,

I agree with you. For a horseplayer, it is way smarter to keep quiet if you have a special insight or information that you believe is more accurate.

For a seller of information, it is way smarter to convince the entire pool of potential customers that your figures are better.

Personally, I think the customers should be stressing the pluses and minuses of the methdologies and not the individual suspect figures. You want to understand \"why\" the figures are different so you can understand whose thinking and method you prefer.

Unfortunately, doing that is very tedious and complicated to explain. It\'s even more difficult for people that haven\'t made figures to understand unless it\'s expressed exceptionally well. Whenever I try it, it isn\'t received very well anywhere. I believe the best approach is somewhere in between the two camps so no matter where I post I am usually pissing someone off and confusing the rest. :-)

[bobphilo]

Michael,

I think what Jerry is saying is that Rags had the Gold Cup so slow that it could have been won by Imperialism with the same figure he earned in his 8th place finish in the BC Sprint.

Bob

[Michael D.]

Bob,
What\'s the point? What does the 6f race have to do with the Gold Cup? I must have missed something here.