A random weapon in the war on terror

A random weapon in the war on terror, from Newsweek.

Airport security personnel at Los Angeles International Airport are now randomly being deployed; the thinking behind it is that if there's no pattern in where the security people are, it's harder for potential attackers to find a hole in the system that can be exploited.

This is the idea of a "mixed strategy" in game theory. There are various things that each of the two parties can do, and the payoffs to each player in each situation are known. It often turns out that the optimal strategy for one player might be to do one thing sixty percent of the time, say, and another thing forty percent of the time -- but that doesn't mean they should do some sort of strict alternation. Rather, they should choose which thing to do at each time at random.

The canonical example of this, I think, is pitch choice in baseball, where the pitcher chooses his pitches at random and the batter chooses what he will expect (and therefore swing at) also at random. But I'm not sure to what extent I believe it actually applies there because the strategy is different on each count. (And it would be incredibly difficult to study. You can maybe get information on what kind of pitch the pitcher pitched. But how could you know what the batter expected?)

The theory of mixed strategies, though, assumes that both parties have perfect information. So it's not exactly the same thing; the terrorists and the airport security don't have such information. Even if they think they know where there's a hole, how do they know they just haven't been watching long enough? But the idea seems reasonable; security people develop patterns, and if you can avoid creating those patterns you can avoid having them exploited.

P.S. I had to mention baseball because the Phillies won today, and are therefore in the playoffs. And the Mets lost; they have to go home now. I was at the Phillies game. See a picture of a sad Mets fan from the New York Times. At one point the odds of the Mets not making the playoffs were 500 to 1, according to Baseball Prospectus. But that's actually not the worst collapse ever; the '95 Angels beat 8800-to-1 odds. (The '64 Phillies are only tenth, according to that list.)