by Mark R. Waser (originally appeared Dec. 10, 2012 at Transhumanity.Net)
You’re appearing on the “hottest new game show” Money to Burn. You’ll be playing two rounds of a game theory classic against the host with a typical “Money to Burn” twist. If you can earn more than $100, you get to move on to the next round.
One of the current problems in (and arguably with) game theory is the so-called “centipede game” (Rosenthal 1981) – where the so-called logical/rational/correct/”optimal” strategy produces virtually the worst possible result. While there are many variants of the centipede game, they all take the form of players alternately choosing between increasing the pot and continuing the game –or– “cashing out” by taking the majority of the pot and ending the game. The “problem” posed by the game is Continue reading