Cognitive Psychology: The Science of How We Think

Here's a question that's bothered evolutionary psychologists for decades: why do humans keep making overconfident decisions when those decisions routinely backfire? Failed businesses. Reckless risk-taking. Terrible judgment calls. The answer, according to new research from the University of Bath and the London School of Economics, might be that our overconfidence isn't a bug at all. It's a feature.

The study, published in Psychological Review (DOI: 10.1037/rev0000644), proposes that human overconfidence functions as a costly evolutionary signal—much like a peacock's tail. And those costs? They're precisely what make the signal credible.

Professor Chris Dawson from Bath's School of Management puts it bluntly: "Overconfidence is strikingly common and often costly. It contributes to failed businesses, excessive risk-taking and poor decisions. But the real puzzle is why evolution hasn't eliminated it. Our research suggests that those costs are exactly what makes it useful."

The Handicap Principle: Why Costs Make Confidence Credible

Think about a peacock. Its tail is heavy. It's energy-intensive. It screams "eat me" to every predator nearby. Yet precisely because carrying it is so expensive, it tells potential mates something valuable: I'm healthy enough to survive despite this ridiculous burden.

Human self-belief works similarly, according to the researchers. People who hold inflated views of their abilities make mistakes—but those mistakes tend to be less damaging for high-ability individuals than for low-ability ones. A skilled entrepreneur who overestimates her chances might lose some capital, but she won't be blindsided by catastrophic failure. A less capable person making the same overconfident leap? That's where the real damage happens.

So the system self-selects. Only genuinely capable people can sustain high self-belief without suffering severe consequences. The cost differential ensures that bold claims remain informative signals of true capability.

"If there were no costs associated with overconfidence, everyone could simply inflate their claims as much as possible, and self-belief would become meaningless," Dawson explained. "What keeps it credible is that holding inflated self-beliefs is less costly for the more able."

Confidence opens doors. But once those doors open, it's real ability that determines who succeeds. That's why the signal works. The more confident are more likely to gain influence, achieve higher status, and access valuable opportunities—from leadership roles to promotions to relationships.

Triversian Self-Deception: Believing Your Own Hype

The research builds on a famous 1976 theory by evolutionary biologist Robert Trivers, who argued that self-deception evolved because it helps us deceive others more effectively. Conscious bluffing often fails—you can't hide the nervousness, the vocal strain, the micro-expressions. But if you genuinely believe your own pitch? Those tells disappear. You become genuinely persuasive.

Trivers showed that truly believing what you say makes you more persuasive. But his theory left an unresolved question: if overconfidence is so widespread, why don't people just assume all confident claims are exaggerated? Why don't we just discount them as cheap talk?

The new research answers that: self-beliefs are discounted, but unless you're overconfident, you'll be underestimated. It's precisely because excessive self-belief carries a price that it can carry information. Everything hangs together because only those who genuinely have something to offer can afford to make the boldest claims.

David de Meza from LSE's Department of Management puts it this way: "An unresolved question has been why people don't simply assume that confident claims are exaggerated. Our research answers that: self-beliefs are discounted, but unless you are overconfident, you will be underestimated."

Loss Aversion as a Hidden Handbrake

Here's where things get interesting. The study identifies loss aversion—our tendency to fear losses more than we value gains—not as an isolated cognitive flaw, but as a co-evolved safeguard. It acts as a hidden handbrake on overconfidence.

"People talk confidently but act cautiously," de Meza observed. "This combination lets you project a high-status image while hidden caution doesn't completely alleviate the costs, but it keeps you from taking fatal risks."

Crucially, this caution often remains hidden. It doesn't undermine the confident image individuals present to others. Overconfidence and loss aversion are, in the researchers' words, "a well-matched pair."

The external megaphone projects status, deters competitors, and secures opportunities. The internal handbrake curbs the physical dangers through covert caution. Together, they allow individuals to project high-status confidence publicly while avoiding catastrophic failure privately.

The Gender Gap in Overconfidence

The signaling logic also explains why men tend to be more overconfident than women. Historically, the sexes faced different mating incentives. While men prioritized highly visible physical cues of fertility, women prioritized traits like status, commitment, and resource acquisition. Because these qualities are harder to observe directly, they must be signaled.

"Overconfidence evolved as a credible way for men to signal this underlying capability," Dawson said.

It's not that women lack confidence. It's that the evolutionary pressures differed, and overconfidence became a more valuable signaling tool for men trying to communicate hard-to-observe traits.

Why Eliminating These "Biases" Could Backfire

The findings challenge a common assumption: that cognitive biases should be eliminated wherever possible. The researchers argue that interventions aimed at removing overconfidence or loss aversion entirely could inadvertently do more harm than good.

Without overconfidence, people may struggle to persuade others or seize opportunities. Without loss aversion, they may take excessive risks. Kahneman's advice to eliminate both biases, the researchers suggest, is poorly founded.

"The findings challenge the common assumption that biases should be eliminated wherever possible," the study notes. "Interventions aimed at removing overconfidence and loss aversion entirely could inadvertently do more harm than good."

From this perspective, these "biases" are symbiotic—the payoff to agents from an integrated set of biases is higher than would be the case in their absence. They're not isolated glitches in the cognitive machinery. They're co-evolved features working in concert.

Sources

How Overconfidence and Loss Aversion Co-Evolved — University of Bath, August 7, 2026. Original research: "Talking the Talk, Not Walking the Walk: The Coevolution of Overconfidence and Loss Aversion" by Chris Dawson and David de Meza, Psychological Review, DOI: 10.1037/rev0000644.

cognitive psychology: the science of how we think

Cognitive Psychology: The Science of How We Think — The Evolutionary Logic Behind Human Overconfidence