Run an Axelrod-style round-robin tournament: every strategy plays every other strategy. See which approach dominates when cooperation, retaliation, and forgiveness compete head-to-head.
In 1980, political scientist Robert Axelrod invited game theorists to submit strategies for a computer tournament of the iterated Prisoner's Dilemma. Fourteen strategies competed in a round-robin. The winner? Tit-for-Tat, submitted by Anatol Rapoport — the simplest strategy in the competition. It cooperates on the first move, then simply mirrors whatever the opponent did last.
Axelrod ran a second tournament with 63 entries, many specifically designed to beat Tit-for-Tat. It won again. The result sent shockwaves through economics, biology, and political science: nice strategies (those that never defect first) dominated.
In a single-shot Prisoner's Dilemma, defection is the rational choice. But when the game repeats, cooperation can emerge because players can punish defectors and reward cooperators. Axelrod identified four properties of successful strategies:
Cooperation depends on what game theorists call the “shadow of the future” — the likelihood that players will interact again. When the future casts a long shadow (many rounds, high probability of continuation), cooperation is sustainable because defectors face long-term punishment. When the shadow is short (few rounds, one-shot interactions), defection becomes rational.
Try changing the number of rounds above: with just 10 rounds, end-game effects can shift rankings. With 1000 rounds, the long shadow of the future makes cooperative strategies even more dominant.
These same strategic principles — cooperation, retaliation, forgiveness — drive the AI in Tactiko. Every turn is a simultaneous decision where you must anticipate your opponent. Will they press forward or hold back? Pass or shoot? The dilemma never ends.