From a young age most people are told to be rational: weigh costs and benefits, pick the option that pays off best. But a new study in Proceedings of the National Academy of Sciences (PNAS) by Marta C. Couto, Fernando P. Santos (Socially Intelligent Artificial Systems group at the University of Amsterdam), and Christian Hilbe at the Interdisciplinary Transformation University in Linz suggests that in social situations, people who are a bit less precise – a bit more ‘noisy’ – can sometimes come out ahead.

Studying how people learn in social situations

Using mathematical and computational models from evolutionary game theory, the researchers studied how individuals learn to act in strategic situations where their outcomes depend on what others do as well, such as negotiating, sharing tasks and deciding whether to help others. In these models, each individual has a sensitivity to outcomes: some strongly gravitate towards strategies that seem to work best, while others learn more erratically and sometimes stick with suboptimal choices.

Most theories and models assume this sensitivity is the same for everyone and fixed over time. The new work breaks both assumptions. The team asked: what happens when people differ in how sharply they react to rewards – and when that very trait can evolve?

The surprise: high sensitivity is not always the winning approach. To see why, the researchers explored classic ‘games’ that capture social dilemmas.

To donate or not to donate? That’s the question

In a donation game, one person can help another by paying a cost – for example, spending time or resources so someone else benefits. Helping is good for the other person but always costly for you in the short term. In this kind of strict, competitive setting, highly sensitive learners quickly realise that holding back pays off for them: they donate less and earn more. As this gives a clear edge over being less sensitive, an evolutionary arms race in short-term gains emerges, even if the group as a whole ends up worse off.

The office kitchen problem

Things flip in a snowdrift game, illustrated by a shared office kitchen. Everyone likes it clean, but each hope someone else will scrub the sink. If nobody cleans, everyone loses. In this kind of situation, being less sensitive can become a strategic asset. Those who care a bit less about the immediate payoff clean less often, nudging more sensitive colleagues to do the dirty work. Similar phenomena are known in psychology as ‘strategic incompetence’ and in biology as the ‘red‑king effect’: moving slower, or caring less about small gains, can shift the burden to others.

Different games, different outcomes

Over the long run, these short-term advantages play out very differently. In donation games, the harsh, competitive setting keeps rewarding sharper and sharper learners, so over time people become increasingly focused on immediate gains. In most snowdrift games, sensitivity also increases at first but then settles at a moderate level: beyond that point, being even more ‘rational’ no longer pays off. A third type, coordination games, behaves differently again: here the population can split, with some individuals evolving to be highly sensitive and others remaining more relaxed, so that both learning styles coexist side by side.

Imperfection is not always a flaw

The broader lesson: noisy, imperfect decision‑making is not always a flaw. In a world full of interdependent choices, sometimes slowing down, caring a bit less about every small gain, or tolerating ‘sloppiness’ can be the smarter long‑term strategy.