The conventional POV is (as Master Ludi said) is to just bet the
overlays. As Master Zero said, the correct answer must be some form of
Kelly.
While both of these statements are true, they are fraught with problems.
In order to
know who the
overlays are, two things must be know about the horses being considered:
1) The probabilities
2) The odds
Since we have neither one, by very definition, there can be no true
known overlays.
The formula for
Kelly is, simply put,:
ADVANTAGE / ODDS
Again, we have neither of those.
When I say, "WE" have neither, I mean those of us with "normal" resources. (I include myself within that group.)
Given a large enough and properly defined database, along with an remarkable analysis engine, reasonably good estimates can be made of both probabilities and odds.
The "Whales" have such resources.
Does that mean everything is hopeless?
No. But it does mean that we must take a direction that does not find us colliding with the whales in every race and still find those "overlays" that Masters Zero and Ludi correctly spoke of.
My Approach
My take on this is that there are 3 things necessary to win
at the races. I am certainly not saying that this is the only way to do it. (That would be both arrogant and presumptuous.)
Neither am I saying that there is nobody among "us little people" making probabilities and properly estimating odds to such a degree that profitability is possible.
What I am saying is that I am simply not smart enough, educated enough, or resource-laden enough, to compete with the whales.
Note: My belief is that if I attempt to "do what the whales are doing," that will put me in direct competition with them, a battle I will most surely lose.
I would include in that endeavor any effort that attempts to use a large database to estimate probabilities on horses with some type of weighting scheme.
So, here is my approach.
(I've posted most of this on PA in the past, BTW.)
1. Pick good contenders.
2. Determine "value" among those contenders.
3. Determine if the race itself has value.
1. Pick good contenders.
Actually, this means "Pick good non-contenders."
Since most people who pick contenders believe the answer is to be found in total winners, this differs greatly from what most people do.
The goal is to be able to say, "When I say a horse is a non-contender, he doesn't win even close to what the public thinks."
More specifically, my goal is to pick non-contenders that will consistently lose about 35-60% of the money wagered by their backers.
(Contenders are those left in after the N/Cs are removed.)
2. Determine "value" among those contenders.
How does one do that if they are not making
probabilities?
Well, the answer is that I am making probabilities, albeit not very good ones. LOL
My plan of attack - and it is relatively easy to duplicate (WHAT?) - is to assign relative probabilities to all of the horses in the race and then penalize the N/Cs 90% of their probabilities.
The impact is such that a 2/1 favorite who has been deemed a "non-contender" might be assigned a ridiculously low probability, like 4%, or something else that is insanely low.
Doesn't really matter. Once I have determined who the contenders are, my goal is primarily to determine who the best bets among them are.
Please note that at no time do I find an ODDS LINE to have value. I always make a
probability that will total to 100% for the field.
About that "WHAT?"
I emphasize that word because most people would say that making a probability is not EASY. It is precisely easy.
It only becomes difficult if you need an ACCURATE probability.
So, what good is an inaccurate probability?
Remember, all we need is something that produces RELATIVELY GOOD probabilities BETWEEN OUR CONTENDERS.
More #2: Value of Contenders
The purpose of the probabilities is to use them in conjunction with SOME PROJECTION OF FINAL ODDS to determine which of the contenders are THEORETICAL GOOD BETS.
Those "good bets" are marked as HORSE TO PLAY.
Note: Final Odds Projection is actually not very difficult. (WHAT?)
Seriously, it isn't. What I do is use several factors weighted together, but you don't have to. Just take the odds at 0, 1 or 2 minutes to post as your estimate. (For this purpose, the best of those would be the first flash of "0" minutes to post, but it really won't matter much.)
I suppose that last sentence could use its own "what?" but this post is already way long. LOL
3. Determine if the race itself has value.
This is where the rubber meets the road, boys and girls, and the answer lies in a relatively simple multi-step formula. (I'm sure that I could make it look all mathematical and s**t but instead I will teach the principles and the steps.
The Principle
1a. The greater the percentage of pool represented by the horses we are going to bet, the less money we should bet in the race.
1b. The greater the percentage of pool represented by the horses we are NOT going to bet, the MORE money we should bet in the race.
2a. The greater the percentage of pool represented by the horses we call "contenders," the less money we should bet in the race.
2b. The greater the percentage of pool represented by the horses we call "NON-CONTENDERS," the MORE money we should bet in the race.
(The important principles 1b and 2b.)
Steps
1. Determine percent of pool NOT BET. ("PctNotBet")
2. Determine percent of pool NOT CONTENDER. ("PctNotCont")
RaceBet= $100 x PctNotBet x PctNotCont
Two Examples
Example 1
You are betting 40% of the pool, and your contenders make up 70% of the pool.
PctNotBet=60%
PctNotCont=30%
$100 x 60% x 30% = $18.
Example 2
You are betting 30% of the pool, and your contenders make up 48% of the pool.
PctNotBet=70%
PctNotCont=52%
$100 x 70% x 52% = $36
My experience has been that the higher the RaceBetAmt the more profitable the race is. Conversely, the smaller the RaceBetAmt amount, the lower the profit.
There is a tipping point at which the races produce a loss.
In my own handicapping, I have found that the tipping point to be $30. Wagers below $30 lose money, and the lower the bet size goes, the more the lose.
SUMMARY
What happens is that you wind up betting more money on the races where you've eliminated the greatest part of the pool.
The hardest part of this entire process is step #1 (picking good non-contenders). The other two steps are really pretty easy.
Perhaps this will provide you with some ideas.
Best Regards,
Dave Schwartz
