Sheets Computer Model Odds line for the ROTW

[derby1592]

Below is an odds-line that was developed using a sheets-based, \"Monte Carlo\" computer model. The primary data input are TG figs.

The odds shown next to each horse represents the model\'s estimate of the \"break even\" odds for each horse. If the horse goes off at higher odds than shown below, then that horse is an considered an overlay according to the model and vice versa.

The model indicates that Transcendental is a big overlay and the obvious key in the race.
The other 4 contenders are Shiny Band, Raging Fever, Mystic Lady and Critical Eye. Based on the morning line, Shiny Band, Raging Fever and Mystic Lady are likely underlays while Critical Eye is a likely overlay.

I have also included some of the relevant \"break even\" $2 exacta prices produced by the model.

Good luck to all.

Chris

****
Win Odds Line:

Hope\'s Expectation, 107/1
Shiny Band, 4.0
Critical Eye, 5.8
Mystic Lady, 6.6
Happily Unbridled, 23/1
Too Scarlet, 60/1
Two Item Limit, 53/1
Raging Fever, 4.8
Transcendental, 2.8

\"Break Even\" $2 Exacta Payoffs

Trans/Shiny,  $37
Trans/Eye,  $39
Trans/Mystic,  $52
Trans/Fever,  $31

Shiny/Trans, $55
Eye/Trans, $71
Mystic/Trans, $85
Fever/Trans, $59

[kev]

Throw out HOPE,TOO S., and RAGING F.

box all others for a tri. and make TRANS your key horse for W/P slots and for the ex\'s.

[tread]

Happily Unbridled looks good here, look at the way Unbridled horses improve from 3 to 4.  That small top last out was a good sign, look for further improvement today at a price.  Good luck to all...

[derby1592]

For what it\'s worth, here is an observation.

I see that Happily Unbridled is popular with the few who have posted about the ROTW and that is understandable. She has the strongest pattern of all the horses entered in the race and is thus an obvious horse for a sheets player to zero in on. Therein lies the problem.

One thing that I have definitely noticed while working with the sheets computer model is that horses like Happily Unbridled (very strong pattern but 2 or 3 points or more slower than several other horses entered in the race) tend to be way overbet. I am pretty sure you can attribute this to the influence of the sheets and condition handicappers and I would imagine that years ago (back in the days of TGJB\'s hot streak) that these horses used to go off at very generous odds and were probably even good plays. It looks like those days are long gone. Happily Unbridled has the best chance to move forward but, if you crunch all the numbers, she is still not that likely to win because there is an even better chance that at least one of the other faster horses (there are 5 significantly faster horses in this race) will run a good race and beat her even if she does improve.

That is why the model rates her as having only an outside shot to win, which is higher than her 10-1 morning line would imply.

Such horses tend to be better values in the show position because the strong patterns imply that they are unlikely to run a bad race. In fact, according to the model, Happily Unbridled has as good of a chance to run 3rd (actually finish exactly 3rd) as all but Critical Eye and Raging Fever.

An interesting contrast is that a horse that is almost as fast as all the others (maybe a point or so slower) with a fairly strong (but not real strong) pattern (e.g., Transcendental) often is underbet.

That is why building an odds-line is so helpful. The best bet is not always the most likely winner or the horse with the strongest pattern. Long term success hinges on betting the best \"value\" not on cashing the most tickets or being the sharpest \"condition\" handicapper.

Good luck to all.

Chris

P.S. I would imagine now that Happily Unbridled will win for fun at long odds. But keep in mind, that almost any time you cash a bet you are lucky (i.e., very few horses have a better than 50% chance to win a particular race). The results of a single race only tell you who happened to be lucky on that particular race. Good horseplayers can only be defined by long term success.

[bj]

*** An observation regarding your odds model ,  the ranking of the odds it produces (from most likely to least likely winner ) is perhaps more important than the actual odds line it produces .
Your model had Transcendental as the most likely winner , the public had her as the fourth choice - this clear divergence points her out as an obvious overlay .
Her performance yesterday certainly proved the validity of that ranking .
The actual odds however were a bit troubling . Tke odds line of almost 5-1 on Raging Fever seems out of line . I believe a good question every handicapper who uses an odds line , however derived , should ask themselves is - would i stand on the other side of the counter and lay that price on a horse ?
I hope this doesn\'t appear as nit picking Chris , again it is just my observation of yesterday\'s and other races you have posted with your model . You are obviously an intelligent and articulate person who happens to be a highly skilled handicapper .
The mathematics involved in your odds model is well beyond my scope .
I use a rudimentary spread sheet analysis to project a race day TG number based upon recent TG numbers at today\'s surface and distance - adjusting for today\'s weight and projected ground loss . It keeps me from overlooking fast horses .
I thought your observations on Happily Unbridled were very insightful.
Ironically i think sheet players themselves are \"conditioned\" to sometimes outmoded thinking in term of condition analysis .
How else did Charismatic and War Emblen get away at the prices they did ?
According to the NY Times interview , Ernie Dahlman doesn\'t believe in the \"bounce\" , and
he says he is the master of the obvious . Well you don\'t get much more of an obvious play then the last out fastest TG number at the same surface and distance .
I look forward to your future posts . bj

[tegger]

Bryan & Chris,

Bryan makes a great point about whether you would have willing to give a patron 5-1 on Raging Fever.  The horse certainly seemed to be by far the logical favorite with a line of 0\'s and a one.  I always look forward to you posting your line and admire the work.  I think there may be variables that you weight too heavily.  The Derby is an excellent example - you took the unorthodox position of penalizing horses for early speed.  If you had left the speed bias out you would have had War Emblem as a much bigger overlay.

Your points about line vs pure fast numbers is dead on.  A horse with a bad looking line that has run 3 ones and two x\'s is much more likely to win a race then a horse with a 6,4,2 or a 5,5,3,3,2 pattern (last number is most recent number in my example).  I usually only improve horses when there is a clear reason (i.e., a horse running for the third time in a maiden that has a line that says poor start, hustled to lead, angled in and tired last furlong).  I guess what I\'m saying is that I love the model and wish I was smart enough and industrious enough to do one of my own but since that won\'t happen I would like to know a little more about some of the variables and how you weight them.

Thanks,

Tegger

[derby1592]

BJ and Tegger,

Thanks for the comments and questions.

Keep in mind that this model is a work in progress. Overall, the model\'s odds-line has been fairly accurate (e.g., a projected 5-1 shot wins about as often as a 5-1 shot should) but I am sure that it is far from perfect and produces some occasional bad numbers. My fellow collaborator and I (actually, he has done most the work) continue to make enhancements as we do more research and test things out. That being said, I agree with you that Raging Fever seemed more likely to win than the odds-line would indicate.

Of course, after the race the winner almost always looks obvious.  

From the model\'s point of view (and mine for the most part since, to some extent, the model just quantifies my own opinion of each horse), Raging Fever was vulnerable. At the weights, there were 4 other horses just about as fast going into the race. She appears to be a better sprinter than router and she also appears to be a need-the-lead type (at least going into the race she did) and she was stretching out an extra 1/16 of a mile after carving out fast fractions under similar conditions and tiring. She is fast and consistent but she had run a long string of consecutive good races and looked like she might be tailing off a bit.

If, she did rate, she was likely to lose some ground from her outside post although this was not that much of a factor given the 1-turn nature of the race.

Finally, add to all that a few other slower horses in the race with an outside shot to run a big race such as Happily Unbridled, Two Item Limit and Too Scarlet.

If you toss all that into the hopper and crunch the numbers, it\'s not that surprising to see Raging Fever rated at a little under 5-1. Maybe it was a little high but she was still likely an underlay at the post-time odds.

Of course, she did win the race but it was hard to anticipate that she would rate off the pace (at least it was for me) and slow down from a 1:09.4 6f time in her last race over the same surface to something closer to 1:10.4. Even so, the hard-charging Transcendental was beaten less than a length and probably ran a faster figure and that was just one possible race scenario. In fact, a very favorable scenario for Raging Fever. Now imagine that she goes straight to the lead as she typically does (a more likely scenario). Do you think she would still have won?

What if Shiny Band had paired her last or Mystic Lady or Critical Eye had run back to their tops? They did not but there certainly was a chance that they might have done so.

Finally, what if Raging Fever simply ran an off race? Obviously, she did not but there was a chance that she might.

The Monte Carlo model runs the race thousands of times to see who wins under all sorts of possible scenarios by assigning different probabilities to the various possible outcomes. It is almost impossible to factor all of these \"conditional\" probabilities into an accurate odds line without the help of a computer and sometimes the results are not that intuitive.

This model is still a work in progress but I think we are definitely heading in the right direction. We have had some short term success with it but only time will tell if we can use it to beat this game in the long run, which is all that really counts.

Chris

[Two Bucks]

I\'ve been following this odds line each week and I don\'t think it\'s ever picked an overlay winner. My dog could do a better job of making an odds line.

[mandown]

Scott,

Yet another pithy observation from yourself. Your generosity of spirit is to be envied. The only think I ask is that you refer to your \'significant other\' with less disdain.

[jaho]

I have to say, from what I\'ve seen the \"model\" is a joke.  It bears no relation to probability of winning, significantly underrates winners, and seems like a useless tool.  I also seriously suspect there is no computer model at all, but that this is just another of the crazies out here, like Jimmy from California who supposedly makes up his own numbers.  Oh,well, if you like to hear yourself talk, so be it.