How to bet on sports and the Super Bowl using game theory

The 2010 Super Bowl matchup is set. The Indianapolis Colts will face the New Orleans Saints. Already there is much speculation about who is the favorite, which quarterback will perform, which defense or offense will show up, and how these dome teams will play in an outdoor stadium.

In fact, none of these things really matter. It is possible to be ignorant and yet make a sure-fire winning gamble, both mathematically and practically.

My goal is not really to gamble. I neither gamble regularly nor do I encourage it. But I wish to illustrate an important point of gambling.

One may recall that gambling is an application of probability. In fact, it was a gambling wager in 1654 that led to the mathematical development of probability theory. The historical details are captured in letters between Blaise Pascal and Pierre de Fermat.

The idea then is the same as it it now: it’s the odds that matter. If you can play the odds, you can win. Here is how to do it in a way similar to the odds line set by Las Vegas bookies.

And yes, this works for almost any event betting on sports, politics, or otherwise.

How to make a guaranteed winning bet

Think about two people who are die-hard fans of their respective teams. Their beliefs can be written in probability terms, but keep in mind this level of precision doesn’t really matter. So let’s get started.

John believes the Colts will win with probability 3/4.

On the other hand, Michael believes the Saints will win with probability 3/5.

Both John and Michael enjoy friendly bets. And they are willing to bet as long as they think it’s a fair deal–that is, as long as the bet has a positive expectation.

Now here’s the fun part. Given these facts, it is possible to design a sure-fire bet that makes you money. It doesn’t matter which team actually wins–you will get money coming to you!

The idea is to make sure your loss for an event is always less than the gain.

Here is a specific way you can construct the bet:

–You bet with John that you will pay him $4 if the Colts win and he will pay you $5 if they lose. John will take this bet because his expected value is positive: (3/4) 4 + (1/4) * -5 = $1.75

–At the same time you bet with Michael. You bet with Michael that you will pay him $4 if the Saints win, and he will pay you $5 if the Saints lose. Michael will want to take this bet because his expected value is positive: (3/5) 4 + (2/5) * -5 = $0.40

The genius part is that both John and Michael believe they are making good bets. They will want to take these bets because their expected winnings are positive. You aren’t doing anything to pressure them or conceal the facts.

So how do you fare? The amazing part is that you net $1 regardless of who wins the game!

See this for yourself:

–If the Colts win, you win $5 from Michael but pay $4 to John. The net is you win $1.

–If the Saints win, you win $5 from John but pay $4 to Michael. Again, the net is you win $1.

You can obviously optimize the bets so you gain maximally, and you can increase the bet in multiples as long as both parties agree.

In financial terms, you have capitalized on differing beliefs to create an arbitrage opportunity. This is similar to what intelligent investors could do when mutual funds and stock options were in their infancy.

This is also something like what Las Vegas bookies do: they set a betting line to encourage both sides of a bet to have even money, and they pay out the winners a share from the pool of the losers’ money. This doesn’t always work in practice, but it is usually profitable over time.

In your case, it might be safer to hide your scheme, or John and Michael might not bet with you out of spite.

Reference:

Su, Francis E., et al. “Sure Betting on Different Beliefs.” Mudd Math Fun Facts. <http://www.math.hmc.edu/funfacts>.

Discussion questions:

1. Above, John has a winning bet. So does Michael. In the long-run, that should mean both make money. But you too are making money for sure. How is it possible that everyone is making money? This is a betting paradox–how to resolve it?

2. You can design a winning bet as long as the beliefs are different. Construct a winning bet if Alice believes an event will happen with probability p and Bob believes it happens with q > p, and both will accept gambles that have positive expectations.

3. Research project: how do professional bookmakers make money on sports bets? Consider various types of bets like proposition or parlays.

The 2010 Superbowl matchup is set. The Indianapolis Colts will face the New Orleans Saints. Already there is much speculation about who is the favorite, which quarterback will perform, which defense or offense will show up, and how these dome teams will play in an outdoor stadium.

In fact, none of these things really matter. It is possible to make a sure-fire winning gamble, both mathematically and practically.

My goal is not really to gamble. I neither gamble nor do I encourage it. But I wish to illustrate an important point of gambling.

One may recall that gambling is an exercise in probability. In fact, you may recall it was a gambling wager in 1654 that led to the mathematical development of probability theory. The historical details are captured in letters between Blaise Pascal and Pierre de Fermat.

The idea then, as is now, is that it is the odds that matter. And odds are nothing more than beliefs which have interesting properties. And below I’ll teach you how you can capitalize on these beliefs.

How to win in gambling

I have two friends that are die-hard fans of the two teams in the Superbowl.

John thinks the Colts will win with probability 6/8.



Share this post:

| More

Previous post:

Next post:



  • BP

    Good point, I learned of this loop-hole from a European friend many years back when he would bet on soccer games between Italy, France and Germany using different online gaming sites, picking out those where the odds leaving him a small margin to make some pocket money regardless whichever team wins.

    But I didn’t understand your comment on “mutual funds and stock options were in their infancy.” Can you elaborate using an example?

    BTW, I enjoyed reading your blogs as you often relate to everyday situations in life.

  • Jon

    IS this more common in europe then? We have known and used this for years now, using the betting exchanges online.

  • Scott

    1. The paradox is that their beliefs cannot represent actual odds. It cannot be the case that one team has a 75% chance of winning and the other has a 60% chance of winning.

    Given the number and complexity of the variables involved, it probably wouldn’t be hard to find people able to justifying contradictory outcomes, and this strategy is to exploit that.

    2. Processing….

    3. Aren’t these called Dutch books?

  • Hari

    1) Thats not true. They both believe they have winning bets. But, at least one of them is wrong. (They might both be wrong if they each win half the time). You can see this a few ways:
    - Conservation of money. Imagine they both have a million dollars to start with. After each round of betting, the total amount of money much remain the same. Since I end up with a positive change in money, at least one of them must lose in the end.
    - Contradictory probabilities. If Pr(not A) = 1 – Pr(A). However, their probabilities don’t match this way, so at least one of them is wrong.
    - Pretend John is right. In the long run, he wins 1.75 each round. Also, Michael loses 1.75 each round in the long run.

    2) Thinking.

  • http://www.simoleonsense.com Miguel

    Hi Presh,

    Already my favorite post of the year. Then again I’m a value inestor and this type of thinking really makes sense to me. Thanks so much

    Miguel

  • Pingback: Simoleon Sense » Blog Archive » How to bet on sports and the Super Bowl using game theory (and always win!)

  • Dilemma

    The dilemma is that you can’t make a guaranteed profit unless you are 100% sure both Michael and John take up your offer, and you can’t be 100% sure of this.

  • http://www.mindyourdecisions.com/blog/ Presh Talwalkar

    BP: The mutual fund and stock options comment was in regards to mispricings in the market, or pure arbitrage opportunities. There were times when a fund was priced cheaper than the sum of its stock holdings! Those days are long over…

  • Sauron

    Hey, it’s Prof. Su! Hell, it’s Harvey Mudd! I am currently in attendance, in case anyone is wondering why I’m so excited to see this on one of my favorite sites.

    Though, for that matter, you should all watch the Colbert Report tonight. Another Harvey Mudd math professor is going to be on!

  • http://www.brelip.com Duo

    1 – well it is zero sum, so it can’t be positive expectation for everyone. I think it is generally true that players are not always rational when they place bets for such a game. People sometime would bet despite perceiving themselves having an edge.





Previous post:

Next post:

Other posts you may enjoy reading: