In a shared office kitchen, everyone agrees the space should be clean—but nobody wants to be the one wielding the mop. This simple domestic tension reveals something profound about human behavior that researchers at the Max Planck Institute for Evolutionary Biology have only recently begun to understand: sometimes, acting erratic or less sensitive to immediate gains can be a strategic masterstroke.
A new study by Marta C. Couto, Fernando P. Santos, and Christian Hilbe, published in the Proceedings of the National Academy of Sciences, overturns a common assumption about how we learn and decide. For decades, researchers modeled decision-making as though everyone responds equally to success—the more a strategy pays off, the more likely people are to use it. But Couto and her colleagues at the University of Amsterdam and Austria's Interdisciplinary Transformation University ask a different question: what if being less sensitive to your own gains, learning in a muddier and more erratic way, actually wins in the long run?
The researchers used mathematical models and computer simulations to test this idea across different strategic scenarios. They focused on "outcome sensitivity"—essentially, how much people care when they succeed or fail at something. Highly sensitive people quickly abandon losing strategies and latch onto winners. Less sensitive people wander around making seemingly suboptimal choices. The conventional wisdom says this sloppiness is always a weakness.
But the snowdrift game tells a different story. Picture that office kitchen again. Everyone benefits from cleanliness, but everyone also benefits more if someone else does the work. In such a situation, individuals who care less about the immediate reward of a clean space paradoxically end up winning. They clean less, which nudges the more sensitive people—who care deeply about results—to do the cleaning instead. The less sensitive people get a clean kitchen for free. Psychologists call this "strategic incompetence," and evolutionary biologists have long noticed similar patterns in nature.
When the researchers ran simulations to see how outcome sensitivity itself might evolve over time, the results diverged dramatically depending on the game being played. In donation games, where people decide whether to help others at a cost to themselves, sensitivity tends to climb ever higher. But in snowdrift-type scenarios, sensitivity climbs initially and then stops at a finite level—becoming hyper-aware of gains no longer helps.
Most striking is what happens in coordination games, where players need to align their choices. The simulations revealed populations splitting into two groups: some individuals became increasingly sensitive to outcomes, while others drifted toward being less sensitive. Rather than converging on a single strategy, diverse behavioral profiles persisted side by side, each viable in its own way.
The implications stretch far beyond game theory. Couto notes that "noisy or erratic behavior does not need to be a by-product of cognitive constraints. Instead, it can serve as a means to gain long-term strategic advantages." In other words, what we might dismiss as human imperfection—our inconsistency, our occasional indifference to immediate rewards—may actually be adaptive wisdom honed by evolution. The messiness of human decision-making might not be a bug. Sometimes, it's a feature.