Axelrod’s second tournament tested whether cooperation could prevail after contestants had learned from the first competition. Its 63 strategies produced a pointed result: Anatol Rapoport’s Tit for Tat won again, while the exceptionally forgiving Tit for Two Tats, which would have won the first tournament had it been entered, fell well down the rankings. The reversal made the tournament more than a contest of programs: it illustrated that strategic success depends on the other strategies present, and that cooperation can flourish without being the same as indiscriminate forgiveness.

Design and changed conditions

Axelrod received 62 entries and added Random, bringing the field to 63. Unlike the first tournament, in which each game lasted exactly 200 moves, the second left the number of moves open. Scores could still be expressed as a percentage of an always-cooperate benchmark, but calculating that benchmark was now more complicated because the game length was not fixed.

The uncertain duration mattered to the logic of repeated cooperation. If players know the exact final round, backward reasoning makes that last round resemble a one-off Prisoner’s Dilemma, encouraging defection; the same reasoning can then unravel cooperation in each preceding round. An unpredictable endpoint instead preserves the possibility that present conduct will shape future responses. The tournament’s open-ended games thus reflected the importance of an uncertain future to strategies built on reciprocity.

A field divided by the first tournament

Contestants had been given the first tournament’s results, including Axelrod’s account of why nice and forgiving strategies had performed well. Some took the lesson directly and submitted strategies of that kind. John Maynard Smith entered Tit for Two Tats, which tolerates two consecutive defections before retaliating. Others anticipated that their informed competitors would choose cooperative strategies and submitted nastier programs intended to exploit them.

That split tested more than whether contestants could imitate a successful strategy. Each school made assumptions about the behavior of the other: cooperators expected the benefits of niceness and forgiveness to persist, while exploiters hoped to take advantage of those qualities once they became common. The tournament’s outcome therefore depended on the composition of the field, not on a strategy evaluated in isolation.

Results and the limit of forgiveness

Tit for Tat, submitted by Anatol Rapoport, won with 96 percent of the benchmark score. Nice strategies generally outperformed nasty ones: all but one of the top 15 entries were nice, and all but one of the bottom 15 were nasty. The overall pattern supported the first tournament’s lesson that cooperative conduct could prosper even when competitors had incentives to defect.

But Tit for Two Tats did not repeat its hypothetical success from the first tournament. The second field contained subtler nasty strategies able to prey on a strategy that was forgiving without qualification. Its poor relative standing did not erase the value of forgiveness; it showed that the result changes with the kinds of opponents present. The contrast between tournaments thus cautioned against treating any strategy’s ranking as independent of its competitors.

Significance beyond the contest

The tournament made the difference between a winning strategy in one particular field and a generally stable strategy explicit. Its results prompted the question of whether there was a less arbitrary way to identify a best strategy, leading back to the concept of an evolutionarily stable strategy: a policy that cannot be bettered by an alternative when most of the population already adopts it. A tournament ranking records performance against its submitted opponents; evolutionary stability asks whether a strategy can resist invasion by a different one.

The broader argument is that cooperation does not require a world in which every participant is unconditionally altruistic. In repeated interactions, the prospect of future encounters can make reciprocal cooperation advantageous, while vulnerability to exploitation places limits on how forgiving a strategy can be. The second tournament’s paired findings—Tit for Tat’s victory and Tit for Two Tats’s decline—capture that balance: niceness did well, but success remained conditional on the strategic company it kept.