Stephen Wolfram applies ruliological methods to iterated game theory by systematically enumerating all possible strategies represented as programs (finite state machines, cellular automata, Turing machines) and having them compete against each other. Rather than studying hand-picked strategies like tit-for-tat, the investigation covers the entire space of possible programs. Key findings include: larger programs can outmaneuver smaller ones by having specialized sub-strategies; adaptive evolution reliably finds winning strategies but produces mechanisms that resist simple description; for cellular automata, winning strategies tend toward simpler behavior; and computational irreducibility means there is no shortcut to predicting competition outcomes other than running the programs. The work covers match-or-not and prisoner's dilemma games, cross-type competitions (FSMs vs. cellular automata vs. Turing machines), and co-evolutionary dynamics where both competitors evolve simultaneously.
