Magic: The Gathering can simulate a computer using its card rules.

How a card game can imitate a computer

The 2019 result was not that an ordinary match of Magic: The Gathering secretly runs like a laptop. Researchers showed that, with a carefully prepared set of cards and game state, the game’s rules can carry out a computation with no fixed limit on how long it runs.

The key is Turing completeness: a system is Turing complete if it can, in principle, simulate any general-purpose computer, given enough time and working memory. The proof encodes information in the state of the game, then uses card effects and the rules for resolving them to update that information step by step. A sequence of turns can therefore act like a program.

The match becomes a computation

The researchers built a setup that simulates a simple abstract computer. Its internal state is represented by pieces of the game—such as cards and their positions—and specific interactions trigger the next computational step. The arrangement is deliberately engineered: this is a proof about what the rules permit, not a claim that players commonly assemble such a board in regular play.

The crucial consequence is that deciding how the constructed game will end can be as hard as determining whether the simulated program eventually stops. If it does, the game reaches one designated outcome; if it keeps running, the computation does too. This connects a seemingly bounded contest to a classic, unsolvable problem in computer science: no general method can correctly predict whether every possible program will halt.

A complexity result, not a ranking

Calling Magic the “most computationally complex” game goes beyond what the result establishes. Turing completeness is a powerful threshold, not a universal leaderboard: it says the rules can express arbitrary computation under suitable conditions, not that every match is unusually difficult or that the game is uniquely complex.

What makes the result memorable is the source of that power. Interacting card effects, persistent game objects, and sequences of rule-driven actions can form a tiny computer—one whose processor is the rules themselves.