Mark Cramer
C & X Report

C&X 43
CONTENTS: Editorial: Bumhood. DRF Speed Rating Plus Track Variant. On Overachievers and the DRF Speed Rating Plus Track Variant. The Jockey Factor. A New Perspective on Automatic Bets. Viagra in the Breeding Barn. On Doping and Unsound Horses. Information Overload. The Trainer Factor: Kevin’s Research. On Speed Figs and Trainer Stats. The Real Edge. The Last Word: Questions on Polytrack.
EDITORIAL: SOCIAL DARWINISM IN THE LAND OF EXOTICS
Not long ago when we posted our Florida Derby day stakes weekend, I was able to pick the winner and second place finishers in the filly stakes race, in order, and neither of them were favorites. I received congratulatory e-mails, and I was elevated to the status of hero. Even failing in the other two races, the profit for the day was significant.
How fleeting is glory. For both the Kentucky Derby and Preakness, where the favorite was recognized as legit but not recommended as a win bet, in order to conserve my hero status, I needed my “exotic inclusion” to finish second. In large fields I ended up with both horses in third: Denis of Cork at 27-1 in the Derby and Icabod Crane at 22-1 in the Preakness.
I took my case before the judge. I said, “Your Honor, getting a 27-1 to finish third in a large field is great handicapping, and coming back with a 22-1 in third in a 13-horse field is nothing to sneer at.”
“Cramer, you have nothing to show for it. You didn’t tell your readers to put the ALL in the second slot in the tri, with your exotic inclusion in third.”
“How could I?” your Honor. “There were too many horses in those fields.”
“That’s the point. You lost. You’re a Bum with a Capital B.”
“But the intention, your Honor, is to give the readers some insight to complement their own handicapping. I certainly provided that insight. Some of them must have collected by throwing in my exotic inclusions.”
“Cramer, not good enough. It’s Social Darwinism out there in pari-mutuel land, good handicappers get eaten alive because they had this or that flaw. You could solve 67 parts of a 68-part puzzle and you’d still qualify as a Bum in the Land of Exotics.”
Tomorrow’s another card of racing. Life is good.
DRF SPEED RATING PLUS TRACK VARIANT
Not a very sexy title, but this method landed me a cold exacta (see next article in my response to Ken).
Since that successful pick on the C&X website, I had another chance to use the DRF Speed Rating plus Track Variant in the Kentucky Derby. Had I applied it to the letter of the law, the trifecta was achievable. However I was not yet at the stage of my investigation to be in a position to use it automatically, and so I let certain prejudices (call them “interpretations” interfere.)
One of those interpretations was probably correct: that the DRF track variant for the Illinois Derby was falsely inflated. That would have devalued Recapturethglory’s 119.
The second interpretation was dead wrong. I stubbornly refused to consider the 110 of Eight Belles because she was a filly, even though she had previous earned a 111 and a 110 (third and fourth races back).
Once I had decided not to use the DRF Speed Rating plus Track Variant in a fundamentalist way (by obstinately refusing to accept the numbers on Eight Belles), I opened myself up to a fatal error. Furthermore, I established the GAP in the “curve” in the wrong place, accepting 104 and above instead of 109 and above. Let’s look at the results of the Derby by DRF Speed Rating plus Track Variant;
RECAPTURETHEGLORY, the questionable 119, finished 5th out of 20: not bad.
DENIS OF CORK, 118 in his next-to-last race (I threw out the clunker in his top line because of the superb workout reports) finished 3rd at 27-1. (He was one of my two “exotic inclusions”. Not bad.
BIG BROWN had 109 in the Fla Derby but also earned a 113 in his previous race, and he was not at his top in either because of hoof problems. You already know that he won.
EIGHT BELLES earned a 110 in her most recent line, and she had earned a 111 and a 110 in her third and fourth races back, suggesting that the 110 was legit.
Only one other horse could have been used, and that was Z HUMOR with his questionable 114 in the Illinois Derby, five points less than Recapturetheglory. Z HUMOR had earned 95 in his two previous races, and this was one reason why I felt the numbers for the Illinois Derby were falsely inflated.
Meanwhile, other horses I liked had far lower DRF ratings. COLONEL JOHN had a 104. I sincerely believed he would move up in his first race on the dirt, because of his bullet work at CD. PYRO didn’t even break the 100 mark in his two superb wins at FG. GAYEGO had a 106 and Z FORTUNE a 105. So there was a GAP of three points between the 109 of BIG BROWN and the 106 of GALLEGO, whereas there was no gap in my 104-and-above cutoff point.
Yes, my friends, this is all after the fact. But this is the type of analysis that must be done, even if painfully so. In the context of my exacta score in the filly race on Florida Derby day, also based on DRF Speed Rating plus Track Variant, and previous research that is only now expanding, I can form a hypothesis that the DRF Speed Rating plus Track Variant has become a valuable underused tool in the past performances.
In studying my own thought process (we should all do this after both wins and losses) I discover that the DRF Speed Rating plus Track Variant caused me to resuscitate Denis of Cork from my eliminated horses and make him an exotic inclusion, but that I did not follow through on the same resuscitation process when it came to Eight Belles.
My own lesson in horseplayer psychology is that I can now be a believer in this particular piece of information, and place it ahead of other more used info from the DRF pps.
ON OVERACHIEVERS AND DRF SPEED RATING PLUS TRACK VARIANT
Letter from Ken:
Mark, great, great call on this race today with 3 winning and 6 running second. What a deal. Made my day. I am trying to learn here. I keep looking at SHE’S ALL ELTISH and trying to understand why you thought she was the most likely winner. Maybe I am missing something obvious, but I am sure it would be a GREAT education if you would share your analysis with me. Ken
Ken,
I really appreciate that you benefited from the information I gave.
Please understand that no matter how good a handicapper may be, he is
wrong more often than he is right.
Let me retrace the steps in my reasoning.
(1) First, as I noted in the posting, Robbie's Gal caught my attention
as an exotic inclusion, because she was an overachiever, finishing
much better than she was expected to do (3rd at 17-1 and 2nd at 33-1),
and switching to the rider of her last victory. Strangely, the race
would not have interested me if it had not been for my taking note of this filly.
(2) Then, the two expected favorites looked very vulnerable. Both
Highest Class and Backseat Rhythm were come-from-behinders that often
fail to get up on time. Also, both had serious trainer concerns: BR's
trainer was ice cold at 4% for the year and HC's trainer, one of the
best, was not good with layoffs, and this filly had been absent nearly
two months (and had not fired fresh as a debut horse). So the two faves looked vulnerable, another good reason to stay focused on the race.
(3) With three year olds I always prefer an improving horse to ones
that have earned reputations because they were "picking up pieces"
against better horses.
(4) Once I saw a good in-the-money possibility at high odds, coupled with
vulnerable favorites, I simply looked for a "comparable", a term they
use in real estate. The comparable here was that four horses in the
field showed flat-mile races at GP. It's embarrassing to show how
simplistic I am, but with straight comparables I often use the DRF
speed rating-plus-track variant (because most people use the Beyer
figs). It's not that one is better than the other but one is overused
and the other underused.
She's All Eltish 86+19 = 105
Robbie's Gal 77+27 = 104
Halo Hollie 84+16 = 100
Overandabeauty 83-16 = 99
and, not entirely comparable because it was from a different track, but
Highest Class also had a flat-dirt-mile race at Fairgrounds with:
89-07 = 96
and Highest Class did not even have a race over the GP surface which
gave yet another reason for his being vulnerable.
(5) As the extra feather in the cap, Castro, who had won two straight
on Halo Hollie was switching off that one and preferring She's All
Eltish.
(6) There was a mild pattern match, since She's All Eltish had won
going turf to dirt and the trainer had 39% wins with this maneuver.
Still, what really attracted me to the race was the ITM chances of the overachiever Robbie's Gal, and that led me to finding the winner.
Hope that provides something to chew on. As you can see, here I’ve analyzed my thought process when it was right, and for the Derby I analyzed it when it was wrong. Knowing which thoughts take place in which sequence is knowing how one’s own handicapping mind functions: the handicapping equivalent of psychoanalysis.
Mark

THE JOCKEY FACTOR
Conventional wisdom tells us that the trainer factor eclipses the jockey factor. Everywhere we find trainer specialty statistics. Rider stats, however, are considered secondary, rarely touted, and, it appears, underused.
We’ve come a long way since the rider factor was emblazoned among the keys to understanding the races. Today we all recognize the trainer as the dominant factor. But jockey info can still be valuable.
In preparing this issue, I went to my usually early meet method of tabulating longshot trainer standings, in which I would write down every winner at $12.00 and up, and then identify the trends. I was confident that I would have some useful information for this issue of C&X.
Luckily, I let my research speak for itself. This time around, beginning of HOL and AQU, no New York or California trainer is particularly dominant in the longshot payoff range. In California, Rafael Becerra became my trainer hypothesis, scoring with three longshots (if one can include an 11.60 payoff. This could eventually lead to something but it remains at the hypothesis stage. Three thousand miles east, there was Randi Persaud, starting out Belmont with a longshot binge but at this writing unable to sustain it.
What I have noticed, mainly on the east coast, is that trainers seem to have two-win longshot binges on the same card. Linda Rice did it, both times with Alan Garcia in the saddle. The never-give-up T. Turner was able to produce two big ones in a single afternoon.
But in general, a remarkable variety of trainers have made cameo appearances in the longshot parade, none of them monopolizing the festivities.
For riders, however, it’s another story. In SoCal we have the 20-year-old Dominican apprentice, Joel Rosario, appearing on numerous occasions in the winners’ circle with horses that did not figure in the pps. Mr. Rosario has achieved the feat for all different trainers: Martin Jones, Polanco, Sise, Abrams, and Greeley. Both Solis and the perennial Martin Pedroza (who never gets the best horses) are also riding ahead of their odds.
Back east, Columbus-born Channing Hill has “pulled a Rosario”, with four longshot winners in the first two weeks of Belmont, all paying more than $25.00, and for four different trainers, none of them leaders in the standings. Apprentice Sebastian Morales has also ridden better than his odds, as has Alan Garcia, C; Velazquez, R. Maragh and JJ Castellano.
As with the west coast, the east coast longshot standings show an amazing parity among many trainers, most of whom are not household names.
Thus, at least this one time, the dominant-jockey factor has replaced the dominant-trainer factor.
Thus far, no concrete jockey research method has emerged. But if you are playing pick 3s or pick 4s, the oddball throw-in horse that could lead to a toteblaster could be a longshot jockey.
You can continue the longshot rider and trainer standings on your own by noting the characteristic of every winner at $12.00 and up, giving extra points when the winners pay more than $20.

A NEW PERSPECTIVE ON AUTOMATIC BETS
An economic crisis has befallen us, worse than the crash of ’29 and the subprime crash combined. It is the Richard Matlow crash. I just crunched out his 5-year stat. Thanks to a lower percentage of winners with first-time starters (9% instead of the usual 17%), Matlow has dipped slightly below the break-even point. His mean mutuel has remained at the same level.
His number of starters per year has gone up, and I suspect that the attempt to “grow” his stable has been the cause of the decline. Please recall that Matlow first-timers have yielded a profit at least since the late 1980s when I began tracking it.
This could be a lesson in economics: that growth is not always the way to profits, and that sometimes a business has a niche level which shouldn’t be tampered with. This stat could also be a glich. He’s still doing much better than random betting, better than the track take, better than any of the usual “primary factors”, better than “best Beyer fig”, better than most players. But a loss of about 2% is a red flag, and it raises the question of whether or not it’s time to get off.
Another more profound question also emerges: is this proof that there is no such thing as a successful automatic bet?
For the first question, it may be okay to stick with Matlow for the spring Hollywood Park meet, traditionally his best season. If he strikes out during this period, we can get off and still have a longterm flat-bet profit to show for it.
This leads us to the more profound question. Are experts like Barry Meadow right? Will all automatic bets eventually break down?
I believe the original question is unfair. There is a confusion between automatic and permanent. In any financial market there are automatic plays, but no investor ever requires that such plays be permanent. The two guys at Goldman & Sachs who discovered an automatic bet against the sub-prime CDOs never considered that their wager would be permanent. The made billions for their company and came out of it with millions in bonuses.
In the history of business, what business has been permanent? Ford used to be an automatic. Would you bet on Ford today? Entire companies that were enormously successful have folded up their tents when the context has changed.
One of my best ever automatic bets was first-time or second-time on the grass for a high Tomlinson turf-rating horse when the rest of the field was comprised of proven losers on the grass or horses with low turf Tomlinson figures. This automatic bet provided me with with regular profits for nearly a decade, often becoming the only wager I would make.
Then the context changed. First, Tomlinson figs appeared in the DRF for everyone to see. Even more important, trainers started to use these figs to spot their horses. Thus, turf maiden fields were no longer full of low Tomlinson fig horses that could be thrown out.
It was not a tough decision to get off. Race after race I was coming up with too many qualifiers and not enough automatic eliminations.
The old cliché is usually correct. It’s easier to know when to get on than when to get off, but when contextual clues abound, and it may be somewhat easier to know when to get off an automatic horse bet than an automatic stock market bet. In the market, the most expert players seem to wait until after the bubble has burst. So the question is not whether one can make automatic plays but how long can they go on.
Therefore, even if the Matlow automatic breaks down, I would not decide to go back and expunge all the profits derived from it.
An automatic bet is not a permanent bet. Even life itself is not an automatic bet. Eventually the sun will die.

ON HEDGING
From Danno:
Hey Mark,
Long time since I sent you an e-mail but I have a handicapping question for
you:
I am an avid believer in developing an odds line and exploiting overlays to
your advantage. I have also effectively adopted your strategy of using a
exacta as place bet and trifecta as show bet. However...
When you have what you feel is an overlay and you bet using the exacta as
place bet and trifecta as show bet aren't you in essence watering down your
odds by hedging your bets? I am sure there is an incremental money
management system one could incorporate to insure they are not turning the
overlay into an underlay but I was interested in your opinion on this.
Thanks...
Danno

Mark responds:
Danno,
Any bet, if you have a true edge, is an overlay. Often a small exacta as place bet pays more than the bigger win bet would have paid. Exacta as place bet and trifecta as show bet are bad bets if they are underlays. Hedging is not a bad idea, so long as both hedge bet and original bet are both overlays.
Good to hear from you.
Mark

VIAGRA IN THE BREEDING BARN
Now that the drugs and racing have become inseparable, how about extending the trend to the breeding barn!
The first horse that could become eligible for this incipient trend is War Emblem, winner of the 2002 Kentucky Derby.
The first drug that could trigger the movement could be Viagra.
You know how medical researchers often remind the public that the fact that an experimental drug has produced results in animals does not mean that similar results can be obtained for humans. In the case of War Emblem, we can advise his Japanese breeders that the fact that the medication works for human beings does not necessarily mean that the success will transfer to horses.
Until recently, War Emblem had proven his fertility by producing 26 sons and daughters who have made it to the track and become successful race horses, with six stakes victories suggesting a high dose of class.
But then, something snapped. War Emblem, supposed to follow in the footsteps of Sunday Silence for the Japanese breeding industry, has now lost much of his interest in the opposite sex. He has not produced a live foal since 2005.
The director of the Shadai Stallion Station laments that “we know he is fertile, but he has no interest in mares.”
Those who know the economics of breeding have said that the stable has already lost as much as $55 million in stud fees.
Shadai seems willing now to sell War Emblem and the price might be right, if you knew that Viagra could get the job done.
There’s one more possibility. What if War Emblem’s problem were shyness? How many of us would have the same level of arousal if we were surrounded by a bunch of leering spectators, some of whom would be manipulating us. Perhaps War Emblem wants to be left alone with his newest belle.
So you vets out there, do you think that Viagra might be the answer?

ON DOPING, UNSOUND HORSES AND OTHER MUSINGS
In a recent column, Andrew Beyer suggests that breakdowns such as the tragedy of Eight Belles may be in part blamed on the lack of soundness of American horses, indirectly resulting from the breeding of fast but unsound horses. Let’s quote:
“The facts are irrefutable. In 1960, the average U.S. racehorse made 11.3 starts per year. The number has fallen almost every year, and now the average U.S. Thoroughbred races a mere 6.3 times per year. Almost every trainer whose career spans the decades will acknowledge that Thoroughbreds aren't as robust as they used to be.
“There are at least two good explanations for this phenomenon. In earlier eras, most people bred horses in order to race them, and they had a stake in the animals' soundness. By contrast, modern commercial breeders produce horses in order to sell them, and if those horses are unsound, they become somebody else's problem. Because buyers want horses with speed, breeders have filled the Thoroughbred species with the genes of fast but unsound horses.
“As this change in the breeding world took place, the sport was allowing the use of painkillers and other medications that are forbidden in most other countries. They allow infirm horses to achieve success, go to stud, and pass on their infirmities to the next generation.”
I have no statistics to back me up but my perception is that European horses break down less frequently than their American counterparts. This may be an optical illusion, distorted by the breakdowns of prominent horses in Triple Crown races, and by the fact that there are so many more tracks racing in America.
However, the first trend mentioned by Beyer seems equally true in Europe. Breeder-owners like Aga Khan readily admit that they love racing but that their bigger thrill is in breeding. That’s why the Aga Khan said, in a press conference following the victory of his now forgotten 3-year-old Dalakani, that he would not race his champion for another year because his primary interest is breeding. I heard the press conference and my perception was that this was his sporting interest rather than a strictly financial one. He is one of the wealthiest people in the world, and racing Dalakani for another year would not have made a dent in his fortune.
But whether for sport or finance, the trend in Europe is also for fewer races per horse, especially with stakes horses, but this does not seem to be the result of a lack of soundness in horses. In the USA, I suspect that the trend of racing horses less frequently has been exacerbated by the prominence of the Sheets among trainers, and the spillover: even trainers who do not read the Sheets are aware of the so-called “bounce factor”, and now prefer racing a fresh horse, and shooting for a win the first time out (sorry for the unintended pun), and then resting the horse.
Beyer’s second point seems more significant. With virtually every horse racing with lasix and so many others using legal painkillers, the probability of passing on infirmities to the next generation is surely increased. In Europe, doping horses is strictly prohibited, and even something as seemingly harmless as lasix is banned. When I was running, I once took lasix as an experiment, just to see if it would make me run faster. I felt no improvement in speed or stamina. On the other hand, today a stiff cup of coffee increases my bicycling speed and endurance. I’m not in competition with anyone, and I don’t consider it cheating, but I can surely admit that a little extra juice increases my chance of enjoying or at least tolerating a steep climb.
In theory I suppose that I prefer a drug-free sports scene, and certainly that is shown to be easily achievable if Euro racing is the shining example. However, in spite of Draconian measures, it has been impossible to prevent doping in the Tour de France, and leaders of the American baseball industry seem not to have a clue about what to do about steroids.
In practice, I’m not so sure. When I bet on an American horse race, I can be reasonably sure that my horse will not suddenly bleed and turn in a clunker. In dope-free France, however, I may choose a live longshot, he may not run a lick, inexplicably, and I will never know if bleeding was the problem.
What is for sure is that, as alluded by Beyer, new fans are not brought into the game when horses break down, and of more long-term significance, when horses race so rarely and their careers are so short, there can be less of a chance to bring in or keep followers of racing, since the potential stars of the game, the horses, do not reappear enough to have a following.
For this reason, in France, attendance and handle at the harness races (the trotters) outdoes the Tbreds. People get to know the horses and these horses continue to perform year after year and even week after week, reliably and resiliently. From the owner’s perspective, it’s cheaper to buy a trotter and the entertainment value is deeper because you get to see him race more frequently.
In America we don’t have the same reaction, with harness racing playing second fiddle to the Tbreds, except perhaps in the New York region.
The system. Harness horses usually race themselves into shape, even if they engage in non-betting “qualifying races”. Horses rarely win after a layoff (the comeback race is preceded by qualifying races, the equivalent of workouts in company). This leads to what I have found to be the most reliable harness system. The rule is simple:
The horse is coming from a qualifying race in which his final time is equal to or faster than the race final times of the other horses.
For the Tbreds I will have to restudy a similar angle: horse shows a workout (4 furlongs or longer) which is equal to or faster than the real race fractions of the other horses in the field. At six furlongs, a horse’s workout can be within a second (5 ticks) of the real race times of the other horses in the field.
For the Tbreds, the workout does not have to be the most recent. The method has worked especially well for maiden races. However, the angle could use a fresh batch of research samples. Any volunteers?
INFORMATION OVERLOAD:
Letter from Don
I have a lot of faith in you and your readership subscribers to C&X. Because of that I have been bidding on some of your out of print books on e-bay and won a couple. My question to you is: how do you digest so much useful information and then recall it when you need to?
I have been trying to get myself into a set routine on how to best cover all the pertinent information--but i sometimes see things after the fact. It doesn't happen as often as it once did-but it still happens. At the moment I first scan for ship ins--then i look at the recent changes in the horse such as blinks, fronts etc. and how the horse has reacted to it. Then into regular handicapping. I was wondering how you first proceed or if you have any thoughts on the subject. I recall your "crib" sheets on proven losers , low % trainers etc. That's the kind of thought process I am referring to.
Thank you for taking the time to continue C&X. I can only partially imagine the huge effort you put forth. And thank you for raising the level of my game.
sincerely
don a.
Mark responds:
Don,
We live in an age of information overload. Look at governments of all stripes: they have all the information available, and they still get it wrong. Look at the major investment banks: ditto. Look at economists: wrong more often than not.
As C&X readers know, I use my “short form” as a basis for approaching a race. Immediately if there are too many horses that are not proven losers at the level or that have trainers win percentages of above 12%, I just move on to the next race, unless it’s a specialty race, or a big Stakes.
Beyond that, or perhaps deeper than that, there is a filtering process in which the mind asks itself a key question: how important is this particular piece of information. Then the mind establishes a hierarchy of importance. In the ideal situation, one, or maximum two pieces of information pop off the page, and they eclipse all others. That’s what I’ve quoted from the book Blink as “thin slicing”.
Everyone else is trying to balance all the information, so we go the other way and extract one piece. I have never been able to get myself to invest in the Vice Fund, for reasons of naïve idealism, but without my money, that fund is more successful than most, even in bad times, because it zooms in on businesses that always do well: those that have to do with vices: arms, tobacco, gambling, and polluting industries. The people who run that fund do not have to consider all the information, because they have sliced out one particular factor that wins.
If there were one factor in contemporary racing that could be sliced out like this, it would be the trainer factor. But I know successful players who do their own thin slicing, extracting “pace” or “class”, etc.
When I do my very best (when handicapping fewer races so I have more time) I scan the pps and wait for something to pop off the page. My memory of all my research and previous play needs to crank up and swirl into something meaningful.
Sorry I can not be more mechanical about this, but there are things here to discuss so fire back any questions that this response has triggered.
mc


CAN A COMPUTER BEAT THE BEST HANDICAPPER?
From Dan M
In the chess world I used to hear the masters scoff
at a computer program beating the World Champion. At the onset of the clash it was sad
how easy it was for the best human to beat a computer. Now of course there is no contest.

I find making money at the races less complex than playing high level chess. So for a computer program
to grind out a profit against the likes of the average horse play is easy to see.

Whenever I talk to horse players they almost all see the game as some sort of art form beyond computer programming.

They often mention music in the same breath as beyond programming. Turns out Mozart had a program for writing
music that used dice, of course he had no computer but he had the program. If I with a purely mechanical method, found
by a computer, can make a profit at the races, then I am sure there are people silently unannounced skimming healthy
profits out of the pools. Just as poker players on the internet are being taken to the cleaners by world class poker programs so
are the bettors of horses. Nerves of steel never second guessing a bet, just apply the Mark Cramer like skills against the masses but with an improvement nothing rattles them. They lose their edge, they stop betting and monitor the betting trends until once again their random high speed genetic search programs find a new edge and bang back in the game.

And I'm not talking about the computer assisted V5, All-in-one, All Ways programs these are not the new breed of AI programs
that are possible. Auto bettors and at best needing only a secretary to phone the bets in but in some cases dialing in themselves through automated accounts where no human takes the bet. Not science fiction because I am absolutely certain that I have the skills
to do this sort of thing and I am hack programmer but is it worth my time and effort, I doubt it.

C&X should propose a match between a top handicapper known to make a profit and stand alone computer program that has only a secretary making the bets and let the human go through the same secretary. The computer monitors the odds and spits out a bet and then is informed as yes or no and the payout. A thousand bets with the computer playing its perceived edge and human playing theirs. The computer will lose the match the first time but next time it will be closer and then the human can't
win anymore and then it will be........

THE TRAINER FACTOR: KEVIN’S RESEARCH
My own database includes information for the ALL-IN-ONE
contenders. I have developed models using the ALL-IN-ONE rating like a
power rating, so that it is one of 3-5 variables that go into my model for
assigning fair odds. The fair odds estimates from these models have greater
wager value than the ALL-IN-ONE "autocompute" odds line. I have found that
the public underbets horses with certain combinations of characteristics
consistently enough that I can make bets for some of these combinations in
the morning, without knowing the post-time odds.

To illustrate a point that you have made repeatedly, trainer win percent is
a highly predictive factor. The table below shows that the top ALL-IN-ONE
contender (autocompute mode) wins about 25% of the races in dirt sprints and
routes (the top 5 contenders win about 83% of the races). However, if the
trainer win percent is 9 or below, the top contender wins only 19.8% of the
time in one of my datasets that covers essentially all non-maiden dirt races
run in North America over several months in late 2007 and early 2008 (15,155
races). If the trainer win percent is 22 or higher, the horse wins about as
often as the post-time favorite (32.6%). As you have pointed out, the sweet
spot seems to be in about the 16-21% range. The win rate goes up with
higher percentage trainers, but these trainers are more heavily bet, so the
ROI flattens or begins to decline, particularly when the trainer wins 25% or
more. Using more specific stats such as those relating to the type of race
or today's jockey are still more predictive.

Non-maiden dirt sprints and routes over several months in late 2007 and
early 2008.

C1 Tr%
#Races
#Winners
%Win
Pay, $
ROI per $
Comments
All
22611
5814
25.7
$37,115.20
0.821
DB as of 5/18/2008
<=9
3356
660
19.7
$5,354.90
0.798
Non-Mdn, Dirt Sp+Rt
>=14
14351
4094
28.5
$23,855.80
0.831

>=16
11898
3474
29.2
$19,794.80
0.832

>=18
9284
2827
30.5
$15,667.30
0.844

>=20
7528
2376
31.6
$12,871.30
0.855

>=22
5375
1780
33.1
$9,262.30
0.862

>=24
3512
1214
34.6
$6,127.20
0.872

>=26
2175
778
35.8
$3,720.60
0.855


Kevin C. Maki, PhD
President
Provident Clinical Research & Consulting, Inc.
Illinois: 489 Taft Ave.; Glen Ellyn, IL 60137
Indiana: 1000 West 1st Street; Bloomington, IN 47403
I hope that you find these comments of interest and helpful. Please feel
free to use the above or any portion of the above in print.

Best,

Kevin
ON SPEED FIGS AND TRAINER STATS
From Dan M
CD Race 8 May 3
On Kentucky Derby Day I made only two bets one of the bets should be finishing right about now.
The other won and paid $28 and I'm annoyed because I had two source of figures ha, ha. It is apparent that
somebody makes their figures with a blender. If you believe pace and speed figures from Bris then the 9
horse Itangaroo had no chance to win with many horses having better pace figures and speed figures, not even
a contender. Then if you went with Equibase figures the horse towers above the rest. I doubt both sources but I do
know the horse won a G1 and and in last two races went by the 4 furlong in 44 2 and 44 3 which is fast compared to the others. One other horse won
a G1 in the race. This horse had out run his odds winning at 11-1 and 26-1, his trainer has a low win rate 4/52
not a significant number if he wins 3 more race he goes from a poor trainer to a an ok trainer. Give me 200 races and then I believe
the trainer stats but the public goes for 52 races as significant so I get strange odds on the real deal.
Now after thinking it over I think it was dumb ass luck.
Two things to learn, speed and pace figures can be nuts and trainer stats on 50 or so races is just too small, maybe with under 5% I might
believe the trainer is not very good.as in really under 10% Generous price because somebody got the track adjustment stuff messed up and some others believe that 50 races
is a statical significant sample. Bullshit!!!
Mark responds:
There are two things here: speed figs and trainer stats. What Dan points out on the speed figs here adds credence to my recent work on the DRF speed rating plus track variant, the most underused piece of information in the whole past performances. Is the DRF plus track variant “better” than the other figs. Surely sometimes, but surely not always. That’s not the point. The point is that this piece of information is virtually ignored by the public so it increases in value.
As for the trainer stats, there are meaningful short-sample stats and there are meaningful long-sample stats, and there are also misleading stats, as we all know. The dilemma is how to “time” these stats. In my work on automatic bets, I have come to a conclusion, as referred to in an upcoming novel I have written: that “automatic” and “permanent” are two very different concepts, and whether or not we choose to believe a stat has much to do with what stage it is in its evolution. It would be unfair to judge “automatic” use of stats on their permanence. No investment, in stock markets, real estate, or any other venue is judged on its permanence. Businesses come and go, depending on their changing context. During the sub-prime period, a few smart players bet against the sub-prime groupthink, and made billions, but these guys also knew that their automatic bet was not at all permanent, and that the context would change and thus nullify the value of their bet. I am working on this subject and will have much more to say, probably in the next issue.
A REAL EDGE
If you could bet a 2/5 horse to show and collect $4.20, would you make the bet?
My wife and I were taking a long and leisurely walk through the Boulogne forest, home of two great race tracks: Longchamp and Auteuil. There are lakes with row boats, an occasional waterfall, and many hidden paths, some of which contain strange looking prostitutes beckoning from behind the trees.
There is also an amusement park. We passed by the front entrance of this attraction and young people of high school or college age were wearing a uniform of the French racing association and passing out 3-by-5 leaflets. We each took one. I stopped and read it:
“This coupon allows you a 1-Euro credit on a 2-Euro bet on Sunday afternoons.” I did a quick calculation. The minimum show payoff in France is 2.20 rather than the 2.10 of the USA. If I made a two Euro show bet on a sure thing but only invested one Euro, I’d get back 2.20 for 1, the equivalent of a $4.40 payoff on an odds-on horse to show!. I had always wanted to take advantage of this edge, since it was frustrating to not be able to play heavily-bet invincible horses.
I glanced into the entrance area of the amusement park. Scanning the scene, I noticed a trash receptacle. It was filled with 1-Euro-off coupons in mint condition. Evidently, amusement-park goers would see the smiling student leafleters and out of the goodness of their hearts, accept the coupons. But the new coupon holders were amusement park goers, the virtual antithesis of horseplayers. Most of them were quickly discarding their clean and fresh coupons in the first trash receptacle before the turnstyles.
This is embarrassing, folks, but I stuck my hand in there and grabbed bundles of coupons. The following day, Sunday, jumpers were going be racing at Auteuil, the favorite track Ernest Hemingway in the 1920s. From my experience, the jumpers can be much more formful than the Tbreds. I had already purchased a Paris-Turf racing form and had it in my backpack.
Once leaving the amusement park, my wife and I found a picnic area. We sat under a cedar tree and ate our sandwiches, which were exceptionally delicious thanks to the crusty 7-grain baguette. After lunch, I opened the form. Two can’t-miss horses popped up off the pages. In the featured fifth race there was the graded stakes winner OR NOIR DE SOMOZA, 18 career races, 12 wins and 17 in the money. The only time he was off the board was when he had fallen on a jump after getting a horseshoe in his face. In the sixth race, another graded stakes event, there was DON LINO, six career races, 4 wins and 1 place, ridden by the dominant jump rider of France, Christophe Pieux.
The next day at the track, betting with these coupons was tedious. You had to go to a window and you had to make 2 Euro bets, and in theory you were only supposed to have one bet coupon, so I had to go from window to window on a crowded Sunday. I put in as many as possible.
The big beautiful strider OR NOIR DE SOMOZA, was named Black Gold because of his shiny black hide. The only way he could be out of the money was if missed a jump, but he looked so rhythmically beautiful in his first two jumps that I knew it was another of his usual days. He tracked the leaders and took the lead before the final two jumps, and then, extended his advantage. Amazingly, thanks to two longshots finishing place and show, he paid the 2.30 to show, almost as much as the win price and the equivalent of 4.60 for a base bet of 2.
In the next race, Pieux had DON LINO about 15 lengths ahead in the early going. At this point, I asked myself, why don’t we always have such a magnificent edge in playing the horses. That’s what it’s all about: getting paid much more than the fair price. Amazingly, DON LINO eventually got caught at the wire, but by only one horse. He returned 2.20 to show, which for me became 4.40 thanks to the coupons.
In the last race there was another Christophe Pieux horse in an 18-horse field. He was not a sure bet. However, there was a powerful factor in his favor. In the morning OTB action he had been 6.5 – 1. In France, at-the-track action is much smarter than OTB play. When the at-the-track bets were averaged in with the OTB action, he was down to 2-1. I still had more coupons. I played him to show. He finished second, paying 3.80 to show, which for me was the equivalent of 7.60.
At regular American races, maybe once a month, there is a horse that should be even money and he’s 4-1. Once in a great while, the public makes a huge mistake. When this happens, it is the equivalent of getting betting discount coupons. Just think! If we only played a race when we knew we had a significant advantage, then we’d have much less action but a much better bottom line.
This is the long-term lesson from the brief miracle of betting with discount coupons. We need to know how to wait. To wait for the rare moments when we have a true edge. The coupons are only symbols here: the betting public, from time to time, will give us a valuable betting discount, no coupon needed.
I still had a fist full of remaining coupons so I went back to Longchamp for the next Sunday. In the first race, there was Coastal Path racing at a 2 mile marthon against only three other horses. He’d beaten all three when needing the race, five seconds slower than his usual time for the distance. He was 5 for 5 lifetime. Five minutes to post he was paying 1/10 in the win pool and 1/10 for place (no usual “placé” wagering, which is show). I used a third of my coupons to place, saving the rest for the fourth race, which I had planned to “buy”. Coastal Path won by daylight. He paid 2.20 to place, but with my coupon, I got back the equivalent of 4.40!
I had planned to use the rest of the coupons dutching (buying) the fourth race, which had five betting interests. I could afford to buy the race because I was getting my wagers at half the cost. I chose this particular race because there was no clear favorite. It looked as if anything could happen. With my huge extra edge, I staggered my win bets with more money on the lowest paying horses, but in way where I would be rewarded more if a longshot came in.
So consider that, in effect, some handicapping intervened. By identifying a chaos race, I told myself that I’d be willing to win less if a favorite won. So I arranged my tickets where I’d only make a 40% profit if either of the top two faves won, a 100% profit if the third fave won, and 150% if either of the two highest shots on the board won. Both had positive angles. One was an informed minority horse (only one public handicapper chose him on top) and the other had a profitable trainer angle.
Please understand, I did not have enough coupons for early retirement, but there was this rare moment in a horseplayer’s life where I could buy the whole race and know that I was going to win money no matter the outcome. It was like going to the track and withdrawing money. Just imagine standing there BEFORE THE RACE and yet knowing when the race was finished that you’d win, no matter what happened.
The informed-minority horse won at 13-1. He was a maiden facing listed stakes winners. I tried to convince myself that I would have played him anyway, but I had no conclusive research about the informed minority in the French form and the other longshot had a legitimate trainer angle.
Let’s return to the lesson, making this can’t-lose bet into a metaphor. I am sure that if any one of us decided to simply wait until we KNEW we had a real edge, our return on investment would rise accordingly. If we waited until we found a horse that should be even money but is paying a hard 5/2 and we can document why he should be even money, then we are gaining the same edge we’d gain by finding real coupons.
THE LAST WORD: QUESTIONS ARISING ON POLYTRACK
Consider the beginning of a Matt Hegarty article in the DRF: Study challenges injury claims
LEXINGTON, Ky. – Data collected over the last six months of 2007 through a uniform injury reporting system has not shown any significant difference in the rates of fatal injuries sustained by horses running on synthetic or dirt surfaces, according to the veterinarian who has compiled the reports.
During a presentation at the Welfare and Safety of the Racehorse Summit on Monday at Keeneland, Dr. Mary Scollay, the Florida state veterinarian, cautioned that the data did not represent a statistically significant set, and it did not include data from the four synthetic tracks in California. Other racing officials said Monday that fatal injuries have declined markedly after the installation of synthetic surfaces at many tracks. But Scollay’s data at least introduces questions regarding the validity of the claim that synthetic surfaces are safer than traditional dirt tracks. Scollay said she was “floored” by the similarity between the numbers of fatalities on dirt and synthetic surfaces.
Mark’s comments:
With the tragic breakdown of Eight Belles in the Kentucky Derby, racing over a track that had been sealed the previous day, I would suggest that research on the injury propensity of regular dirt tracks should be divided into two sets (at least!) one on tracks that have been sealed and the other on unsealed strips. A backup sample would be wet tracks that have not been sealed. This research is entirely doable, given the immense databases that are readily available. At this writing I am not aware of what Churchill maintenance had done to its track in the morning prior to the Kentucky Derby.
Then there is Andrew Beyer’s opinion, referred to above, that the surface begs the question and that it is the widespread use of drugs and the breeding for pure speed that has allowed genetic weaknesses to thrive, gain pedigree value, and be replicated into future generations. Bleeders that would normally not accumulate a stellar career records and thus not pass on their genes to future racing generations of thoroughbreds are turned into non-bleeders, their physical weaknesses masked (and even further masked by shortening their careers for financial reasons) and then new generations of unsound thoroughbreds take to the track, with a greater likelihood of breaking down.
If Beyer is right, then the energy dedicated to artificial racing surfaces is poorly spent. Perhaps in 10 years we will look back at the various types of polytracks as glorified Astroturf. However, some of the best and most conscientious trainers sincerely believe in these surfaces, and want the very best for their horses. The jury is still out.