Cyclical Luck

[DPCondit]

[ QUOTE ]
Poker is a much less complex game than horse racing

[/ QUOTE ]

Didn't you just provide the explanation right there?

If horse racing is much more complex than poker, than there could conceivably also be correspondingly much greater edges as well. Is this not intuitively obvious?

Yes, the rake is enormous, and this was the first thing he ever taught me about horse racing, yet a very good handicapper can substantially beat the rake. This is a fact. If your computer buddies couldn't figure it out, then they weren't looking at the right things, it is more than just number crunching.

Don

[Copernicus]

"If horse racing is much more complex than poker, than there could conceivably also be correspondingly much greater edges as well. Is this not intuitively obvious?"

"complexity" is not quite the right word. I am not referring just to additional considerations, which indeed could lead to greater edges, but to more "unknowables". Poker is a game of incomplete information, but the underlying basis is discrete and unchanging probability distributions. In horse racing the underlying basis is "continuous" and constantly changing. However, this isnt the main problem facing a handicapper, just something that can interfere with those "best laid plains".

Using Bayesian and Neural Nets you can look at "everything" that is available to a handicapper and the techniques will learn what there is to be learned. With electronic past performances available (so there is no element of deciding what to input), there is no issue of looking at the "wrong things".

The problem, as they proved (not "discovered", since this is what is really intuitively obvious) is that the knowable predicting factors that carry heavy weights in the network (ie those that are the best predictors) are also the most obvious and widely known to bettors. If the vast amount of money bet is based on common predictors that account for 90% of racing outcomes you can rarely beat a rake of more than 10%. Even when a more obscure predictor would carry enough weight to influence the prediction, throw in those unknowables and rake becomes even harder to overcome.

To put it a little differently, they had no problem successfully predicting outcomes, the problem is indentifying overlays. The "market" is just too efficient to put in stock trading terms.

That leaves non-public knowledge as the only element that holds enough promise to make betting profitable, and that is an angle, not handicapping.

Andy Beyer, probably the greatest handicapper who ever "went public", makes far more from his books and newsletter than he could ever have made from handicapping.

[DPCondit]

Sorry, you caught me in an ornery mood when I read it yesterday. "Intuitively obvious" may be overstating my position a little.

However, I stand by my position. The only thing they "proved" was that THEY could not beat the races over time, not that it could not be done by someone else.

First off, they are horses, not computer chips. Observing even just how they carry themselves as they are being walked out towards the gate can yield quite a huge amount of information. This is information on how the horse is feeling today, right now. This is not past performance data that can be analyzed on a computer.

You mentioned "non-public knowledge". You got a little warmer with that one, but no bullseye. When do you think the insiders (those that actually have non-public knowledge) get their money in on the race? What does this tell you? Understanding how and when the insiders bet, and using that information to your advantage is not angle shooting, and can yield a large amount of information. Once again, this is not something that will show up when crunching the data on the computer.

Only betting the races with the best overlays.

Knowing how to combine horses properly in Daily Doubles, Trifectas, and Pick Sixes can often yield huge overlays.

Personally, I am better at poker than horse racing, but there are people who beat the horses very substantially long term, and that is a fact.


Don

[Copernicus]

I said a couple of times that I was referring only to handicapping, and we may have a semantic difference. Observing the pre-post/post parade and betting patterns are not, to me, handicapping.

As far as finding overlays in exotics, yes, they exist, but too few and with too great a variance to play without a huge bankroll.

Ive met a lot of people who claim to beat the races consistently, and more people who claim to know someone who does. I've been on both sides, betting and owning, and its difficult enough to make money betting when you are the insider, much less without that information.

I'm not even from Missouri, but I'll believe when I see it, and Ive been hanging around tracks for over 40 years and havent seen it yet. But then, I'm an atheist, too.

[DPCondit]

I also replied with the word "handicapper", lol. I am obviously using the word to have a broader meaning than you suggest.

Yes there is very large variance requiring a large bankroll to properly do the exotics, especially the way he does it.

Anyways, I understand and respect your skepticism, but I have seen what he can do first hand. I have heard his stories of the hundreds of thousands he made on pick sixes. I know his family and have heard their stories and listened to his wife complain repeatedly that he didn't go to the track often enough because he always made so much money when he went (true story). I know friends of mine that also have gone to the track with him on various occasions, they claim (for whatever it is worth) that they have never lost money when they went with him.

I know I haven't convinced you of anything, so be it. I think we must agree to disagree.

Good luck,
Don

[jon_1van]

I have seen a system similar (because it is about Roulette) to this in a book I found at Barnes and Nobles.

The basic outline of the strategy was that you bet on all numbers that form a contiguous portion of the roulette wheel. Ie you bet an extact pie shaped portion of the wheel. Then since your bets are all clumped the prob of your bet winning is actually higher than the fraction of the wheel you placed bets on.

I could believe I read this in a REAL book.

[Moonsugar]

The crazy thing about randomness is that sometimes it doesn't 'seem' random.

This is only a problem with human observation.

[Copernicus]

If the dealer is biased so he lands on that quarter of the wheel more than 1/4 times then they might be right!

Most newspapers have "lottery" columns, that tell you to track prior numbers, and that certain combinations of numbers are more likely to come up than others (not just that they are likely to pay more when they do come up, but that they come up more often!)

Convince someone that you have an easy way to make money, and they will work very hard to make you rich.

[Bozeman]

If he is talking about "The Eudaemonic Pie", it is considerably more complicated than that. They were a group of grad. students who modelled the physics and found that if you note the starting position and speed of the ball (often constant for a particular croupier), as well as some of the properties of that particular wheel (most are not perfectly balanced) you can predict with a reasonable edge which quarter it will land in. They had several observers and a computer/transmitter/reciever in a shoe. They claim to have made some money, but eventually stopped because the computer in those days was not reliable enough in the humid atmosphere of a shoe.

I am sure they made more money with their book,
Craig

[PokerNoob]

Its my understanding from talking to trainers and stablehands that some predicting factors are not knowable to the general handicapper. That skews results enough so that you can't beat the rake. On the other hand, for the puppies, the weights of the predicting factors tend to distribute themselves in such a way that that the degree of confidence in the prediction can't beat the rake either.

Racing is more... chaotic(?) than cards. Given a fair shuffle, the odds for certain hands are a mathematical certainty (over a long enough number of trials).