Axelrod's celebrated Prisoner's Dilemma computer tournaments, published in the early 1980s, were designed to find effective ways of acting in everyday interactions with the strategic properties of the iterated Prisoner's Dilemma game. The winner of both tournaments was tit-for-tat, a program that cooperates on the first round and then, on every subsequent round, copies the co-player's choice from the previous round. This has been interpreted as evidence that tit-for-tat is an effective general-purpose strategy. By re-analyzing data from the first tournament and some more recent data, we provide new results suggesting that the efficacy of tit-for-tat is contingent on the design of the tournament, the criterion used to determine success, and the particular values chosen for the Prisoner's Dilemma payoff matrix. We argue that this places in doubt the generality of the results and the policy implications drawn from them.