First, regarding Jimbo and Janis\' getting together to do the study, I\'m all for it, if it can be done right. That doesn\'t just mean in terms of the study itself, but in terms of the auditing-- after the horses are rated, the ratings should be e-mailed to both Len and me before racing, and the results tabulated on a running basis daily and posted (presumably here, since Len doesn\'t want any part of this). This removes the chance of one side or the other crying foul after the fact. I also think that NoCar Tony might want to get copied, since he will have other ideas about how to do the studies. I think Saratoga is a good meet to use, since there are a lot of shippers, and we can look at track-to-track consistency that way.
My idea is this-- take the last 3 on the same surface (turf or dirt), regardless of trainer changes, time between races, off tracks, etc. Throw out the worst one, average the other two, and you have a rough power rating for the horse. Sure, there are plenty of situations where the methodology will screw one side or the other. But over a decent size sample (300+ races) that should wash out about evenly. Then adjust the ratings for weight to be carried today, and rank the horses for each product in each race. After the race, see who had the winner ranked higher, and that side gets a point.
There are lots of other things you could do (like give an extra point if they had the winner on top), but this is a simple approach that should be indicative. There obviously is some margin of error, and if it comes up 52/48%, nobody should be crowing. But if it comes up 170/130 or so for 300 races, that would be a piece of evidence-- not proof, and we can always do it again or another way (and in fact since both products are sold publicly, anybody can do any study any time they want). But it would be evidence.
The idea here is not to approximate handicapping-- for starters, different players can look at the same data and have different opinions (like Vito thinking Barbaro was a great bet at 6-1 in an evenly matched 20 horse field, and many Ragozin guys disagreeing). The idea is to check the overall accuracy of the figures-- as several people have noted, there are now enormous differences in two sets of figures that in theory are supposed to be measuring the same things, differences that go way beyond small pattern reads, so somebody is doing something wrong. This should give us a clue as to who it is.
So yes, if this is going to be done properly, I\'ll go for the cost of both sheets for Saratoga. Jimbo and Janis won\'t be paying for their data, but there will be some work involved, and if they want to increase the sampling by adding Del Mar for the same time period I\'ll go for that as well.
As far as the comments from the other guys on Ragozin\'s board:
Dave (Soup? Zat you? You\'re very calm, these days)-- I addressed some of your comments above. This is not a betting system, it\'s a figure checking system. And if one set of data is considerably off, pattern reads are irrelevant.
By the way, it\'s also not \"entirely about pattern reads\", to a large degree it is about how fast the horses are. Vito can talk about Barbaro\'s pattern all he wants, but if that horse is 2 points slower going in, it\'s unlikely he bets him. Also, someone who uses our data (head of programming at TVG) called me to tell me about a conversation with someone that uses Ragozin (Amoss) about a race at Keeneland, the Shakertown. Amoss said Atticus Kristy had no shot, he was too slow, coming off a pair of 10\'s. We also had him coming off a pair-- of zeros. Not only did he win (8-1), but the second finisher came out of the same race at Tampa. We both had the winner off a pair, but I think we got it right.
Vito-- yeah, you\'re right. It is of no consequence that Ragozin got the ground for the winner of the Kentucky Derby-- who was on the TV screen every second-- wrong, by 3 paths, and didn\'t correct it when it was pointed out. This is the second time in 3 years they have made a major error (as opposed to a difference in judgment) in the biggest race of the year, and din\'t fix it. Who cares, right?
Raz-- Yes, I know there are handicapping contests. I know because guys using our data have won the biggest one 3 of the last 4 years, and it\'s a contest that is NOT open to the public-- you have to qualify by doing well in other contests. (And by the way, the winner of one the earlier NTRA Nationals, as far as I know the only one to win using Ragozin, has switched to our data).
As for the challenge to Len-- I AM willing to put up my own money, a lot more than I would have to put up in one of the public contests. I want it to be head-to-head, and meaningful-- a large sampling, not 20 races in a contest that depends on being the one to hit the longshots that day. In fact, I am less proud of the winners of those NTRA contests than of the large numbers of TG players that have qualified-- to me that\'s more indicative. The only contests these days where you see large numbers of Ragozin players is the Orleans-- you don\'t have to qualify, and you don\'t have to bet real money, like in some of the contests now. (By the way, you might recall that in 05, despite being wildly outnumbered, our guys did unbelievably well at the Orleans, much better than Ragozin\'s. This year both did poorly).
Indulto-- my eyes glazed over trying to read your fourth paragraph, so you\'re going to have to live without a response.
Addendum-- above I mentioned Atticus Kristy, who came out of grass sorints at Tampa, which we had about 6 points on the scale faster than Ragzoin did. Well, I just cashed the winner of the turf sprint stake at Calder (6th), who was also coming out of turf sprints at Tampa. He paid $27, had 3\'s to 5\'s coming in. Anybody know what he had on Ragozin?
I\'d make some modifications to take the distance into account, you wouldn\'t want to use a fig for a 6f sprinter going 1 1/8 or vice-versa....
John
These things will average out and will affect both parties anyway. I assume it is the relative performance rather than the absolute performance of the services that is being tested. Just my tuppence but if TGJB believes his numbers superior he should seek to use _all_ the horses. ie a 7 runner race yields 21 match bets. This would also increase the data returned for the same amount of sheets consumed.
I know I\'m fairly new to posting, but as far as the comments about possibly adding Del Mar to the study, as an Economist who works in applied Econometrics, and 20+ years handicapping, I\'ll gladly volunteer to do the Del Mar work
tmc-- you should shoot a person-to-person message to Jimbo about this, he and Janis of the Ragozin board are going to do the study.
\"[Jimbo] and Janis are going to do the study\".
Does this mean that Jimbo and JJ are going to monitor the respective figures\' efficacy during the Saratoga meet by identifying the fastest horse and seeing how frequently the fastest horse wins, or that Jimbo and JJ will be making selections based on their reading of the numbers?
I have said this before, about a year and a half ago. To take centuries of breeding, years of training, and a minute or two of athleticism and daring on the part of the jockey, and to distill it all in to one number, well that takes some chutzpah. I think at the time I called it \"a tremendous leap of faith\". That being said (and repeated), I think that the TG guys have taken that single number and used it as the foundation for some pertinent ancillary data. The \"number\" or \"fig\", when combined with the ancillary data, results in a very impressive product.
Let me go a step further. The information which is exchanged on this board, when it doesn\'t dissolve into utter nonsense, is both fascinating and informative. The nonsense even sometimes is gut bustingly funny, and laughter is important.
When TGJB takes Len and the Rag boys to task, I do not think he is trying to establish the superiority of the TG product,but rather to point out that the East Village crew has suffered some lapses in basic approach over the last few years; it becomes I would imagine more infuriating to JB when his criticism is not responded to.
It is not important to me to know if Rag is \"better than\" TG or vice versa. I do not think it is scientifically possible to ever establish this. Maybe it would be found that one service is more effective than the other on certain circuits, on certain surfaces, or even at certain distances. If I really needed to know which product was better, I think I would concern myself with what is called in the newspaper business \"circulation\". Who is selling more of their product? How many owners and trainers are using (and paying for) the respective products? What are the circulation trends? Even this measure of the products might be made difficult by the fact that many professionals are likely using BOTH of the products.
The only true measure of the effectiveness of these products is the long term success or failure of the end users of the product--US. I think that the measurement of that success or failure over the long haul is impossible.
I guess it makes for good debate, and New Yorkers never get tired of this kind of thing. My father and uncle argued for years whether Stage Deli or Carnegie Deli was the best in Manhattan. (Katz\'s on Houston is IMO better than both). As long as there is no method of making any kind of definitive determination, the debate can (and will) continue.
Richie-- Janis and Jimbo are going to do the work in what will be a purely mechanical study, as detailed at the top of this string. Yes, there are many factors that can affect an individual result-- but in the long run they will cancel out for both sides, and faster horses will beat slower horses, as said elsewhere.
There are problems with judging the product by \"results\"-- like, how do you get the results. Getting the ROI of all users of each product would do it, or maybe a long term handicapping contest between Friedman and myself-- I assign you to figure out a way to make either happen.
The other concerns apples and oranges. Ragozin basically gives data away to horsemen-- if you were trying to buy the most valuable horse in the country right now, you could get his sheet for $25. This results in having bragging rights to a lot of \"customers\". We work on a comission basis, controlling the dissemination of the data, and making it (and our opinions) cost more and worth more to our clients. This has resulted in Da Hoss, Victory Gallop, Distorted Humor, Super Frolic, etc.
Yeah, Katz\'s. Where the closest thing to a green vegetable is a french fry. That neighborhood has really changed, if you haven\'t been there recently.
TGJB:
That\'s quite an assignment.
There are only two people who can arrange the mano a mano. I think it would be wonderful to see a joint effort on the part of TG and Rag where your picks and Len\'s picks and the thinking or pattern analysis behind them was produced for a big racing day like Belmont day or Travers day. It would be interesting to see how many times on one racing card the two of you agreed on the fastest horse.
I realize that having mentioned TG, Rag and \"joint effort\" in the same sentence will lead some to accuse me of having partaken of a joint of a different sort before doing this post
Indulto provided an end user efficiency meaurement scenario which I wish I could say was beautiful in its simplicity. It involves teams, multiple race tracks, auditors, kool aid and TV shows.
I am worried about the great Saratoga showdown. I do not know if it is the proper venue to be used for measuring the products\' effectiveness.
1) It is a very small sample.
2) 2YO races, steplechase races, marathon turf races, races taken off turf all make the sample smaller.
3) Yes there are horses converging on Saratoga from all over the world, but many of these animals will be making one or maybe two starts at the meet.
Not to detract from the study that Jimbo and Janis will undertake, but I think a more informative survey would involve a longer meet with less variables. Calder, Penn National, Finger Lakes and Mountaineer all come to mind, but I realize it is not too sexy to monitor long meets at tracks where the winners are running TGs in the high teens.
I haven\'t seen a response to P. Eckhart\'s suggestion. I think it\'s a good one.
Richie and Bit-- it looks like it will be both Del Mar and Saratoga. I saw Eckharts idea, and am waiting for others-- this won\'t be starting for a while.
But a) doing it either the way I originally mentioned or his way or another that makes sense, 600 races is a decent sample, and as Janis has pointed out, if there is a strong trend in either direction, the study could be extended, and b) if we did use Eckharts suggestion, it would just be with the first couple of horses, otherwise you are getting into who runs less bad.
Right now my thoughts are to do the following, after creating a rating for every horse as described above:
1-- Take the top horse in each race by both, do win % and possibly ROI ( latter not really relevant because nobody is saying these are good bets as opposed to fastest horse, but what the hell).
2-- Compare the top horse on each (and possibly others per Eckhart) to see who finishes higher.
3-- Take the winner of the race, and see who had him higher. In case of a tie, go to the second finisher.
Per a post by Indulto yesterday saying that our scale is lower, and figure range tighter-- that will not matter. We will be rating the horses against each other on one set of figures, and seperately against each other on the others, then comparing only the order of the ratings.
Jerry,
Let me know how I can help.
Just curious - since both sites have published data why not use historical data and generate your results ASAP.
You could recap last year\'s SPA meet or the past 5 years for that matter.
NC Tony
You da man.
Saratoga last year, the last 5 years. Any track that the services covered in the last year, the last 5 years. All the data is already stored. Some one out there could figure out a way to feed all this info in and extract some accurate results.
In my next life, I only ask to be educated by the Jesuits.
Unfortunately my feeling (I would love to be wrong) is that neither capo would want this information to be pursued and released without knowing in advance what it would reveal.
Whether or not it\'s practical to use all , that makes sense and would certainly hasten results of compairing differant numbers system\'s . Does anyone know if Beyer is still in ? For logistical purpose\'s , the use of all horse\'s in a race could be restricted to Stakes ...
Tony-- For one,the other guys would never allow it-- they won\'t participate in it the way we\'re going to do it either, which is why I\'m going to have to spring for the cost of a set of Ragozin for 6 weeks at 2 tracks, and comp Jimbo our data for the same period. Also, since Ragozin\'s people wouldn\'t be auditing it, nobody would believe it if we could do it, and TG came out ahead-- unless you are able to send to Friedman the rankings in advance every day he can claim we screwed with the data.
Ahh They don\'t trust you I see. You would think that since it\'s all public record as are the results you couldn\'t screw with the data. What about the archive sections ( I assume RAGS has one) all that is public and free materials now, why not use that info now as a forerunner to the actual excercise itself.
I guess by giving them some time up front now, you are being fair to them by allowing them to get their data inline prior to the SAR meet.
But I hear what you are saying. It\'s best that is on the record too, and we are not all thinking this or assuming why it is they way it is.
NC Tony
PS I\'ve tried to go on the RAGS site but it is so poorly laid out it made it difficult to want to peruse it al all.