I remember when I was in graduate school one of my professors who taught game theory mentioning that every year around this time he would get contacted by the school newspaper asking for strategy tips regarding NCAA Tourney Pools, to which he would respond that these are actually not “games” in the classic manner of game theory, because they were not essentially strategic. I’ve thought a lot about this over the years, and here’s what I make of his claim:
* Game theoretic problems are strategic, in that they are designed to elicit your best move based on your opponent(s) move
* NCAA Pools, in contrast, never have a best strategy other than pick the teams you think are most likely to win, and therefore are _not_ dependent on your opponent’s choices
* Therefore, NCAA Pools are not game theoretic
Now, this is of course not to say that there are not better or worse _strategies_ for winning NCAA Pools. For example, we all know a 10 seed is more likely to beat a 7 seed than a 16 seed is to beat a 1 seed. Moreover, it is possible to design NCAA Pools that are explicitly game theoretic. For example, you could make each round worth a number of points (e.g., Round 1 games are worth 100 points, Round 200 games 200 points, etc.), with the points being split equally among everyone who picked the winner in that game; in this case, your payoff would be a function of the moves made by other players.
But the classic NCAA Tourney pool (1 pt for picking Round 1 games correctly, 2 for Round 2, 4 four Round 3, etc.) should not be considered game theoretic. And yet, I can’t help but wonder whether this apparently simple conclusion might be wrong. Yes, if I can pick all 63 games correctly, then I should do so (and win $1,000,000 on Yahoo while I’m at it). But if I assume that I am going to get some games wrong, then might my strategy differ if I think my co-pool members are more aggressive (e.g., tend to pick more upsets, don’t send four No.1 seeds to the final four) than if I think they are conservaitive (e.g., tend to pick favorites, tend to have a final four compoased of all No. 1 seeds)?
So with all this in mind, I wanted to throw this question out to the readers of The Monkey Cage. Are NCAA Tourney Pools game theoretic at all? Put another way, should my strategy _ever_ depend on what I think other people in the pool are doing?
