Researchers showed that simple games in which rewards and risks change over time can produce far richer strategies than classical game theory predicts. In the Prisoner’s Dilemma, when there are no variable rewards, the model reaches, within a few rounds, the only stable point at which everyone betrays everyone else and all players lose. When rewards change even by small amounts, cooperators and defectors can coexist; when the variation increases further, the defector equilibrium becomes unstable and only cooperators remain.
These results emerge differently in the “chicken” game. When rewards are fixed, the equilibrium in which everyone survives by swerving turns, with a small change, into a population that does not swerve. More noise creates two stable states that alternate between survival and collision.
In the rock-paper-scissors game, under fixed rewards, the strategy does not settle at a single point; the population continuously cycles among the three options. Random changes in rewards can create new stable and unstable points. When the rewards are unequal, cycles emerge in which the probabilities of choosing rock, paper, and scissors change over time in a predictable way.
The study shows that, alongside player behavior, changes in the external environment also strongly influence the best strategy. This approach suggests that the more complex dynamics observed in real life can still be examined using simple game models. The findings were published in Physical Review Letters in 2026 under DOI: 10.1103/3yby-qq2n.
Why it matters
The significance of the study lies in showing that strategic equilibria arise not only from players’ moves against one another, but also from the variability of the conditions in which the game takes place. This demonstrates that classic models built around fixed rewards do not necessarily have to produce single, unchanging outcomes; under the same rules, multiple stable states, cycles, or forms of coexistence may emerge. The findings illustrate for researchers examining game theory and behavioral dynamics together how incorporating environmental fluctuations into a model can change its outcomes. However, the conditions under which these structures observed in simple games can adequately explain complex relationships in real life, and how variable rewards can be measured, remain the main questions raised by the study.