Evolved for Arrogance

Why does nature allow us to lie to ourselves? Humans are consistently and bafflingly overconfident. We consider ourselves more skilled, more in control, and less vulnerable to danger than we really are. You might expect evolution to have weeded out the brawl-starters and the X-Gamers from the gene pool and left our species with a firmer grasp of our own abilities. Yet our arrogance persists.

In a new paper published in Nature, two political scientists say they've figured out the reason. There's no mystery, they say; it's simple math.

The researchers created an evolutionary model in which individuals compete for resources. Every individual has an inherent capability, or strength, that simply represents how likely he or she is to win in a conflict. If an individual seizes a resource, the individual gains fitness. If two individuals try to claim the same resource, they will both pay a cost for fighting, but the stronger individual will win and get the resource.

Of course, if everyone knew exactly how likely they were to win in a fight, there would be no point in fighting. The weaker individual would always hand over the lunch money or drop out of the race, and everyone would go peacefully on their way. But in the model, as in life, there is uncertainty. Individuals decide whether a resource is worth fighting for based on their perception of their opponents' strength, as well as their perception of their own strength. Both are subject to error. Some individuals in the model are consistently overconfident, overestimating their capability, while others are underconfident, and a few are actually correct.

Using their model, the researchers ran many thousands of computer simulations that showed populations evolving over time. They found that their numerically motivated populations, beginning with individuals of various confidence levels, eventually reached a balance. What that balance was, though, depended on their circumstances.

When the ratio of benefits to costs was high--that is, when resources were very valuable and conflict was not too costly--the entire population became overconfident. As long as there was any degree of uncertainty in how individuals perceived each other's strength, it was beneficial for everyone to overvalue themselves.

At medium cost/benefit ratios, where either costs or benefits somewhat outweighed the other, the computerized populations reached a stable mix of overconfident and underconfident individuals. Neither strategy won out; both types of people persisted. In general, the more uncertainty was built into the model, the more extreme individuals' overconfidence or underconfidence became.

When the cost of conflict was high compared with the benefit of gaining a resource, the entire population became underconfident. Without having much to gain from conflict, individuals opted to avoid it.

The authors speculate that humans' tendency toward overconfidence may have evolved because of a high benefit-to-cost ratio in our past. If the resources available to us were valuable enough, and the price of conflict was low enough, our ancestors would have been predicted to evolve a bias toward overconfidence.

Additionally, our level of confidence doesn't need to wait for evolution to change; we can learn from each other and spread attitudes rapidly through a region or culture. The researchers call out bankers and traders, sports teams, armies, and entire nations for learned overconfidence.

Though our species' arrogance may have been evolutionarily helpful, the authors say, the stakes are higher today. We're not throwing stones and spears at each other; we have large-scale conflicts and large-scale weapons. In areas where we feel especially uncertain, we may be even more prone to grandiosity, like the overconfident individuals in the model who gained more confidence when they had less information to go on. When it comes to negotiating with foreign leaders, anticipating natural disasters, or taking a stand on climate change, brains that have evolved for self-confidence could get us in over our heads.


Johnson, D., & Fowler, J. (2011). The evolution of overconfidence Nature, 477 (7364), 317-320 DOI: 10.1038/nature10384