I have the head of data collection for Equibase hounding the photo-finish guys for answers on the beaten length questions, supposed to have them tomorrow. This is actually pretty interesting stuff, and since they changed to digital machinery sometime recently I\'m not going to shoot my mouth off until I know what the situation is.
By coincidence, one of the photo-finish companies had a full page ad in the program on Belmont day. There was no information there either, just one line and a photo.
Thanks for remembering. I\'m looking forward to reading what new information comes out of this.
I\'m getting there, and my head hurts from going back through stuff I hadn\'t looked at or thought about in oh, 20 years or so. First thing is, there is no such thing as a length, except as a unit of time, no matter what it looks like on the photo-- more on that when I get the rest of what I\'m looking for. Dealing with these photo-finish outfits is evidently like dealing with the Kremlin, but we do enough business with Equibase that I should be able to get it.
These photo-finish companies are ridiculous with the way they guard the \"super-secret\" formula to figure out how many lengths a horse was beaten by.
BTW, both time and beaten lengths are available for quarter horses and I think that may be the key to reverse engineering what they are doing. I looked at the relationship briefly last week but came up empty. If you don\'t get any solid answers from them, I\'ll take a harder look when I get a chance.
As I posted before, the Photo Finish Company Employee was solely responsible for determining the margins between each horse on the film strip (which was black and white). They did so arbitrarily, and they were the only word on that subject. Up until I left racing in 2001, this was the procedure at just about every track, and probably still is in place in many smaller tracks today.
Equibase had no oversight authority to these companies (I don\'t know if they have any real oversight today either); the Photo Finish Company\'s contract was with the track ownership, and the Photo company employed a person to run their photo finish equipment. It was this Photo Finish employee which gave out the \"rundown\" as it\'s called to Equibase (and any other chart making outfit), the Horseman\'s Bookkeeper (for payment of purses); sometimes the Mutuels Department, and a few others. After the placing judges detrmined just the order of finish, the Photo person then assigned the margins to each horse.
Bottom line is, up until the year 2001 or so, Finish Margins were determined subjectively by whomever was working the contracted photo company at each track.
The \"margins\" were and are actually a measure of time, which was then translated into inches on a page, and then into \"lengths\"-- but is still a measure of time, not distance. The way they were doing it in the old days greatly created the opportunity for human error, just as the old timing devices for the individual horses (photo finish cameras with moving film) created the possibility of mechanical error. The new machines, in place according to what Equibase has told me at most tracks since the mid to late 90s, in theory solve both problems-- not only are they supposed to be super accurate, but I believe they translate and transfer the data electronically, removing the chance of human error. I\'m still working on this stuff, although nobody seems real happy to be answering questions.
TGJB,Any news from Equibase yet on this issue? Asfufh
That all major sports except ours use PROFESSIONAL and up to date timing systems to record final times and placement. Even my high school has this timing system when we host track meets!!!
When are we going to learn???? Are we ever going to come into the 21st century using this technology...NO because of the political climate at our racetracks employing former jockeys, trainers, agents et all in areas where people with this knowledge of technical education would put the good ol boys out to pasture!
http://www.finishlynx.com/
http://www.amb-it.com/
The answer is, yes and no. Kafka and Heller would love this one. This is what I know as of a conversation yesterday:
Equibase gets its data from the tracks, which in turn have contracts with various photo finish companies-- Equibase has no contracts or leverage with the photo finish companies themselves. The guys we work with at EB do care, and have tried to get answers, but the people at each individual company refuse to answer, except to say that the information is proprietary, they use the same time for a \"length\" at all distances, and that the one they use is absolutely accurate (and I note here that they all may or may not be using the same conversion). Now that EB is aware of the situation they are going to try to at least standardize it, but that will take time.
A \"length\" is a unit of time, as I posted earlier. In response to questions, all companies apparently did agree they are using somewhere around .16 or .17 per length, but would not be more precise than that. We use .16, and it looks like there could be a margin of error of up to somewhere around 5%-- meaning about a half length either way at 10 lengths, or about 1/4 point for a horse beaten 10 lengths at a mile.
Those who know me can probably guess that there were some pretty wild (and funny) exchanges on this subject, like the one where someone said to me \"look, when a running back gains three yards, is it always exactly 3 yards\"? The part of my response which can be posted was, \"but here they KNOW exactly how far it is. And the correct analogy is saying he gained 3 LENGTHS, and when someone says \'what\'s a length\', you say \'can\'t tell you\'\".
Beyerguy and SJU5 have it exactly right. In some ways this industry really is a joke.
That\'s pretty much what I posted earlier: Equibase has no jurisdiction over each Photo Finish Company.
And the idea that beaten lengths is a function of time rather than distance is a dubious assertion. I can imagine it\'s relevant IF a track\'s teletimer is directly integrated into the Photo Finish Equipment. That intergation is MUCHO expensive to do, and a lot of small to medium tracks never did it (and I\'ll bet a lot still haven\'t). I know what I saw with my own eyes for 17 years: races were timed with a hand held stopwatch in many places, even up until I left racing in 2001 (and good that they did because race teletimers malfunction often enough). If someone tries to explain to Jerry, me or anyone else that beaten lengths is a \"function of time\" under these circumstances, then send them my way: I\'ve got a bridge for sale.
Why do beaten lengths matter anyway? If you watch someone like Jerry Bailey ride, beaten lengths as a reflection of a performance\'s quality is questionable at best. Once Bailey sees his horse is getting nothing, he wraps up on them, and they coast home. Shane Sellers, Kent Desormeaux - same thing. If these horses aren\'t at maximum effort from start to finish, how accurate is it to use beaten lengths as a figure making guideline for a horse that\'s not being asked to do anything? What does it reflect?
For those that don\'t know, Equibase is a company formed by The Jockey Club & The TRA for the sole purpose of track ownership of Past Performance data, which then was the exclusive property of the Daily Racing Form. When the DRF came under Rupert Murdoch\'s ownership, fear struck that the DRF would \"disappear\" and tracks would be up a creek with no PP data, since it was all owned by the DRF and published in their paper. So TRA memeber tracks began hiring their own charting crews and went into direct competition with the DRF compiling chart data for races, and for several years you had what could only be a nightmare for figure people: two sets of PP information about the same races every day. Imagine that: the track program with PP data compiled by Equibase chart people, while the DRF PPs in their paper were compiled by their own chart crews. Discrepancies? Uh, many. So much for exactitude.
About 6-8 years ago, the DRF terminated all of it\'s track chart crews and simply bought all it\'s PP data from Equibase (the Equibase crews had a big advantage over the DRF in that they were directly employed by each track, whereas the DRF crews were \"media guests\" that could literally be denied access to any track that chose to deny them - and some were!).
Regardless of whom was compiling charts, the fact is that Photo Finish Company Employees were responsible for determing lengths beaten. It\'s their information to give - or not.
Millenium-- beaten \"lengths\" are a function of time. They have no way of measuring the distance. The photo finish camera takes a series of photos at high speed intervals AT the finish line, the photos are pieced together and presented as a strip, which is calibrated. The time is then measured by counting the calibrations, and expressed as \"lengths\", which as it turns out is not an exact term-- at least wasn\'t until I got involved and started yelling. By the time we get done with this it may be standardized.
Your account of the formation of Equibase is pretty close, except that the tracks approached the DRF first to try to get a reduced version of the pps to put in the program, for which they were willing to pay a royalty. The DRF turned them down, which turned out to be a $300 million mistake-- they were sold for $400 around then, most recently for about $100 now that Equibase has full pps in the program.
As I have said here before, the Equibase data base currently in use is the one that our own George White developed for the Racing Times when he was president there-- he knew EB was on a parallel path, and did a deal to share costs and the data. When the RT went overboard with \"The Old Man\", as George called him, EB had a fully functioning data base, and we became their first corporate client, with George setting up and running the Thoro-Graph data base. Eventually, the DRF figured out they could cut costs by buying the data rather than gathering it themselves, and George\'s little old data base is the only one in the industry.
JB - I understand. Guess I was confused what you meant by \"time\". In this case I assume you mean it\'s the amount of \"time\" it takes a horse to pass through the lens at the finish. Regardless, the expression of distance separating each horse is determined by each Photo Finish Company Employee. One employee in particular at a track I worked NEVER gave margins of 1/4 (neck) or 3/4. It was either a whole length, or a half a legth. That was it; there was no arguing with him.
I knew that The Jockey Club & The TRA tried to buy the DRF Database. I guess the DRF does regret it now. I worked at Beulah Park - it was the first track in the country to use the Equibase Program with the Past Performances. Believe me when I say that the TRA tracks were licking their chops financially too: this gave them all the reason they needed to jack up pricing of a program (the ratio of what the printed costs are to what they charge is enormous).
Tracks make very little in profit beyond things like concessions, parking and admissions. Program sales are monies mostly all profit for each track.
In spite of all this, I still maintain beaten lengths is much ado about nothing. It might matter when horses are \"all in\" for the whole race. Once their \"Millionaire Jockeys\" (as Lukas likes to say to tweak them) wrap up on them, how far they get beat is a weak barometer to measure the quality of an effort. Trainers are the same way - there\'s still the \"give \'em one\" crowd. This year\'s Bluegrass Stakes is a perfect example: Bandini (at the time of the race, April 16th) looked real doubtful to make the Derby on earnings. He needed to run lights-out to be sure he\'d get in. All the others behind him (HL,CA, SK, Consolidator, etc.) already had a Derby starting spot for the asking with their earnings, but needed to get a race in so as not to go into the Derby off 7-8 weeks without a race, which is a guaranteed loser preparation. None of the ones behind Bandini was gonna floor it for $750k in a prep when the $2000k big dance is 3 weeks later and they\'re already in if they choose. And if any trainer operates otherwise, they need to have their licenses revoked.
Doesn\'t beaten lengths HAVE to be a measurement of TIME, and not distance. They appear to be measuring the amount of additional TIME it took for the trailing horses to cross the finish line. There are no beaten lengths, as that would be a snapshot of the field as the WINNER crossed the Finish Line, a measurement that is not used.
So its all about the conversion ratio thats used (time/lengths). As long as you *know* what it is, the actual figure barely matters to an outfit like TG , as they are using FINAL TIME and not lengths to arrive at the figure. They could use .15 .20 .50 , and the figure would be just as accurate.
TGJB, you agree?
Boscar, For maximum accuracy, I think final speed figures have to be computed using the same measurement for beaten lengths as is used in the charts. Otherwise, the final speed figs for the losing horses which(I assume) are currently based on the winner\'s time plus the time associated with beaten lengths would be inaccurate compared to the winner\'s final fig.
Of course, one way to avoid this confusion would be to publish the losers\' final times as measured (very accurately?) by the camera in the charts along with the beaten lengths.
The tracks/equibase could probably do the same(i.e., use cameras) at the internal points of call in order to publish accurate times for all horse\'s pace figures.
Don\'t hold your breath on anything being done by the racing business on these issues. Asfufh
Correct, and I brought up the subject of publishing the times of all the horses Friday. We\'ll see where this all goes.
Jerry,
If you pull this off, you are truly a man among men! :)
When you really think about it, the only way to get the info we horseplayers want is to continue to demand it. Believe me, noone else is going to do it. The tracks sure don\'t care, and neither do horsemen. We are the only ones that can pull it off, and having a serious customer like TGJB lead the fight is a start.
What can we do to help?
I\'m going to be working with Equibase on this, we\'ll see what happens.
I\'m not sure I understand this completely.
Would this help with the problem of horses of different quality finishing at various rates of speed at the end of a race or do they already account for that?
For example, do they accurately capture the difference between top turf and dirt horses when they give us beaten lengths?
Good turf horses routinely run sub 12 second fractions in the last furlong while you rarely see that on dirt. That should somehow be built into the beaten lengths.
Of course if we can get the actual times, then that\'s much better. Good luck with your effort.
In theory, what they are giving us is actual time of each horse, expressed as \"lengths\". I repeat, in theory. Without having more information, which I may or may not be able to get, it\'s hard to know if they are executing correctly.
In other words-- you may see a race where you are sure a horse gets beat 8 lengths, but because of a slow pace they are coming home very fast, and the chart will show him only getting beat 6 because the elapsed time was only 6 times (fill in time value of a \"length\").
Jerry,
I\'m pretty sure Equibase does actual time and actual btn lenghts for Harness Racing Already? I know Sports Eyes puts in the Horses Actual Time,Qtr Splits, Btn Lenghts et al. I believe Equibase is also the official data base for Harness as well. Seems strange to me that they can\'t offer same data for Thoroughbreds as well.
NC Tony
PS How\'s the Golf Game?
The golf game is nonexistant.
I just got off the phone again with my guy at Equibase. This is going to be a process, but now that they are aware of the questions they are definitely moving on it, and they\'re going to bring me in on it as it moves along.
I did find out the following-- the timing process is far more sophisticated than it used to be even 10 years ago, when there was a moving strip of film. They are doing very high tech timing of each horse and the information and images go directly into a computer. Then that super accurate time data is converted into \"lengths\"-- which is where the problem is.
When you make your figures, do you convert beaten lengths, ground loss, and weight to time, and then use that aggregate adjustment to calculate an adjusted time used in assigning figures?
If so, do you use the same number of seconds per length (I think you mentioned 0.16) at all distances?
My overall impression is that the Equibase/photo-finish-company situation, while not as clean as you would like it or as it should be, doesn\'t represent a significant problem for your figures. Is that your take as well?
I\'m sure you weary of discussing this topic, but if people use Equibase data for beaten lengths, how would the other guys have blown the beaten lengths in the 2004 Derby?
Respectfully,
BitPlayer
We convert time, \"lengths\", ground, and weight, into points on our scale. A length has a different value in points at different distances-- about a point per length at 5f, gradually extending to about one point being worth about 2 lengths at 1 1/4 miles.
As I said in an earlier post, based on what we have found out thus far, the degree of \"looseness\" in the figures could be up to about a quarter point at 10 lengths at a mile. By the time the smoke clears on this, everything should very tight.
Ragozin does not deal directly with Equibase. As far as I know they still have a deal with BRIS, which gets the data from Equibase. My guess is that it was a transcription/input error at some point in the process-- somehow they got the 2 length gap between third and fourth in the 04 Derby as 5 lengths. Manual errors can happen to anybody, and happens to us-- but we don\'t make nearly the number of mistakes or ones of that magnitude in big races, and when we make one we admit it, and correct it.
\"and when we make one, we admit it and correct it\"
I can tell all from my personal experience that this statement is true. That is a big reason that TG deserves credibility. Failure to acknowledge mistakes creates a credibility gap and leaves people relying on information from others at a loss.
The speed figures chapter of Six Secrets, which I read windshield wiper style while waiting for a book club meeting to end, has this interesting & marginally related theory from an anonymous contributor: It isn\'t the accuracy of sheets figures which makes sheets players such strong players, but their absolute belief in the figures\' accuracy, which gives them the needed confidence to pull the trigger when it counts. Not sure I totally agree, but certainly food for thought.
Thks to M3 for bringing this interesting subject to everyone\'s attention. I\'ve tried to think of some way of turning this info into an edge at the windows,at least until it\'s changed or more widely known, but couldn\'t think of one. Anyone else come up with anything?
Mall -
Tying this back to an earlier thread, one might term that the \"never-in-doubt-often-in-error\" school of handicapping.
JB Said: \"I did find out the following-- the timing process is far more sophisticated than it used to be even 10 years ago, when there was a moving strip of film. They are doing very high tech timing of each horse and the information and images go directly into a computer. Then that super accurate time data is converted into \"lengths\"-- which is where the problem is.\"
JB, Any progress in getting a definitive answer on the time vs.lengths formula from Equibase etal.
BTW, I assume that the time assigned to lengths in the Equibase charts is the same no matter what the distance of the race. If this is correct, could you explain (probably for the umpteen time) why a TG point is \"worth\" more points the longer the distance of the race. Thanks, Asfufh
Asfufh-- I don\'t expect there to be any progress on the timing issues for a while-- I\'ll check back in with Equibase after Saratoga.
If you want to get an overview of the points/lengths question, think about the average winning margin in quarter horse races, and in 1 1/2 mile races. The longer the race, the more the horses spread out-- a difference in ability translates into more lengths the farther you run.
TGJB, Any word from Equibase on the time/beaten lengths issue? Asfufh
The equibase beaten length issue is going to take a long time. The right guy there is on it, I\'ll ask him again some time after the BC. When I talked to him a couple of weeks ago he did confirm that the differences between the different timing companies were very small, and all very close to what we use. But it should be standardized, and in the end they should be publishing times straight up for each horse.
TGJB, I don\'t get it. What takes so long for Equibase to poll their chart making companies to determine the time they use for each beaten length and then let us handicappers know the results? Asfufh
Asfufh-- they did poll them, but the teletimer companies don\'t work for (have contracts with) Equibase, just with the tracks. Equibase has to get the tracks to apply leverage to get anything done. I was given an off-the-record overview, anything else will take time.
On 9/21/2005, TGJB said;
\"Asfufh-- they did poll them, but the teletimer companies don\'t work for (have contracts with) Equibase, just with the tracks. Equibase has to get the tracks to apply leverage to get anything done. I was given an off-the-record overview, anything else will take time\"
TG, Anything happening on this? I notice that Aqueduct is now posting beaten lengths for all runners about 5 or 10 minutes after a race is run so the time being assigned to each length must be readily available. I don\'t understand what the big secret is. Asfufh
Asfufh-- back in the office after a day at home with a stomach virus (trust me you don\'t want to get this one). Yes it\'s ridiculous, and I\'ll check in with Equibase later this week.
A couple of members of my Yahoo group live in South Africa (not exactly the center of world racing) and yet the charts have the PRECISE times for ALL finishers in 100th\'s of a second. Unfortunately they don\'t take fractional times there (like in Europe) but they seems worlds ahead of us in acuurate finish times of the horses.
Bob