BIG PLAYS AND THE JOB OF THE HANDICAPPER
© Robert Crowne & Associates, December, 2011
"A one year record is totally irrelevant to me. Give me someone’s five-year record or lifetime record that’s what’s really significant. Take a coin and flip it 80 times and you’ll be shocked how often you get 50 tails and 30 heads.” -- Fezzik after winning the Hilton Supercontest twice in a row
I keep reading posts in the various sports-betting forums from people who seem to believe that if a handicapper is any good then his or her big plays must win. Nothing could be further from correct. Handicapping is not about guarantees; it is about probability.
Handicappers do not lose games. Teams lose games. I don’t know a single professional handicapper who goes onto the field or court and plays the game. A team can lose even though the handicapper correctly calculates the probability and predicts a win. The handicapper does not pick absolute, can't-lose teams; he picks teams that he calculates to have a better probability of winning than the posted odds. The farther away from the posted odds, the bigger the play, but that does not guarantee a win. Evaluating a handicapper takes more than a mere calculation of wins and losses.
To accurately evaluate a handicapper based on wins and losses requires looking at a series of between 500 and 1000 selections all in the same subdivision of the same sport and all similarly rated. To evaluate a handicapper in less than 500 to 1000 selections in a sport requires evaluating the methods used by the handicapper. If the methods used should accurately calculate the probabilities, a win. A poster recently wrote, “I don’t care how smart you are, you are not a winning ‘capper.” Such a statement is an oxymoron. If the methodology is correct the wins and losses will eventually reflect it.
As a practical example, let’s take a look at Fezzik. He was the first Hilton contestant to ever win two years in a row. Both years his final record was over 64%. People hailed him as a football genius. Some told me that Fezzik was the living proof that I was wrong when I said that no one could consistently win over 59%. To his everlasting credit, Fezzik himself attributed his high percentages to an incredible amount of luck. Then in the third year Fezzik finished the contest at just below 50%. This year, after 12 weeks, he stands at 45.6%. After winning over 64% on approximately 165 selections, he is now well below a coin flip on the last 130 selections.
So which is if? Is Fezzik a football genius or, in the words of the critic posted above, is Fezzik “not a winning ‘capper?” The answer is that unless Fezzik reveals how he handicaps the NFL, we simply don’t know. The run on the first 160 games, and the run on the next 130 games, and the whole group of 290 games together are simply not sufficient to make a judgment. Come back after another four years of selections and I’ll tell you if Fezzik accurately calculates NFL probabilities. Right now, he’s on schedule to be below the 59% limit on handicapper wins. I suspect he’s neither a genius nor a loser, but somewhere in between.
When speaking about handicapped probability, I always use the same example regarding black balls and white balls in a bag. This time, however, I'm going to change it slightly. Suppose you have three bags, each with 10 balls in it. The first bag has 1 white ball and 9 black balls. The second bag has 3 white balls and 7 black balls. The third bag has 5 white balls and 5 black balls.
Now let's imagine that the posted spreads and money lines on picking a black ball out of the bag are:
Bag 1: -850 (9 black balls and 1 white)
Bag 2: -200 (7 black balls and 3 white)
Bag 3: + 150 (5 black balls and 5 white)
All of the above bets provide an edge, but which play would get the top rating as the BEST BET OF THE DAY? The handicapper's job is to go into each bag, count up the black and white balls, calculate the probability of a black ball being the first one picked, and compare that to the posted odds.
Bag 1 Obviously, in the bag with 9 black balls, the probability is the highest that a black ball will be picked, but that bet is the worst of all of them. The true odds of picking a black ball from that bag are 9-1 in the bettor’s favor, but he or she must lay 8.50-1. That's an edge, but a very small one. If a bettor makes 10 picks from Bag 1, and bets $8.50 to win $1 on black each time, he or she will win $1 nine times, and lose $8.50 once for a mere $0.50 profit on $85 risked.
Bag 2 With Bag 2, the odds of picking a black ball are only 2.33 - 1 in the bettor’s favor. The odds the bettor must lay ARE 2-1. That means that if you bet $2.00 on 10 picks from the bag, you will win $1 seven times for a total of $7, and lose $2.00 three times for a total profit of $1 on an investment of $20.
Bag 3 Even though there is only a 50-50 chance of picking a black ball out of Bag 3, you will get +$1.50 every time you do it. That means if you bet $1 on each of 10 picks you will lose the $1 five times, but you will also win $1.50 five times for a profit of $2.50 on a risk of only $10.
Obviously, if the handicapper is doing his job, the bag with a 50-50 chance of picking a black ball will be his best bet, even though the likelihood that any one pick from that bag being black is the worst.
Note that just because there is only one white ball in Bag #1 does not mean that black MUST win on the first draw. If the posted odds are raised from –850 to –950, the bet on black would have a negative expectation and the bag with the most black balls should not be bet at all. If, however, the odds were reduced to 11-10, the bet would probably become the Best Bet of the Year with a break even of 52.4% measured against a 90% probability of winning..
So how did moving the odds around change the probability that a white ball would be the first one picked, and the handicapper’s Bet of the Year would lose? It didn't. So long as there is a white ball still in the bag, there is always a chance the white ball will be the first one picked out of the bag, and that chance stays the same no matter what the posted odds. You need to make hundreds of picks from the bags before you can determine if the handicapper has accurately calculated the probability of each bag.
The probability of a win from a bag with a fixed number of black and white balls does not change when the posted line changes. Only the quality of the bet will change with a change in the odds or spread. A loss on the first pick of a ball from the bag can occur despite the handicapper being absolutely correct in his calculations. If the first ball picked is replaced into the bag and another pick is made, a loss can occur a second time, and the same thing can happen again a third time in a row if the bettor keeps doing the same thing. Each loss will be an independent event. The balls in the bag for the second pick don’t know that a white ball appeared from the first bag. If there are 7 black balls in a bag of 10 balls, a white ball will be picked a second time in a row with the same frequency as it will be picked the first time.
In sports, a 70% chance of winning is about as good as it ever gets. The likelihood that two games correctly handicapped with a 70% probability of winning is once in every 11 pairs of picks. In other words, it happens. Given that, what does two or three losses on big plays say about the quality of the handicapper? Absolutely nothing.
You can, however, make judgments based on the handicapper’s methodology. For example, a handicapper who opens the bag and counts the balls is likely to be more accurate than a handicapper who says to bet black because black has won 10 times in a row without counting the balls in the bag.
The reason a bettor will be a winner betting on accurately handicapped 70% propositions is not because he will never lose. The reason he will be a winner is that he will have seven wins for every three losses. He will win two in a row 49% of the time, and lose two in a row only 9% of the time. The rest of the time he will split. At the end of the season that spells W-I-N-N-E-R.
In sports betting, there is never a zero chance of a loss, even in fixed games. In fact, if a handicapper accurately calculates the probability that a proposition will win to be a huge 70%, then, at the same time he is telling you that the game will lose three times in every 10 bets, which is once in less than every three bets. If you are foolish enough to over bet a selection just because it is labeled a big play, and then it loses, don’t blame the handicapper. There is a good chance your big loss is your fault for bad money management, not his fault for a bad handicap. Money management is no less important than accurate handicapping.