Why Are We Sometimes More Likely to Succeed When We Are Overconfident?

Psychologists have long noted that humans tend to be overconfident—overestimating their capabilities, their control over events, and their invulnerability to risk (phenomena collectively known as “positive illusions”). Although we might expect an accurate view of ourselves and the world to be the best way to succeed in life, psychologists have suggested that, in fact, positive illusions might be essential to mental health. Without a positive outlook, we lack the ambition, morale, and resolve to try things and gain things. Indeed, positive illusions disappear among people who suffer from depression!

But even if positive illusions help us deal with the personal mental challenges of everyday life, they pose an evolutionary puzzle because when it comes to important interactions with other individuals, other groups, and the world, why would natural selection favor inaccurate assessment? That would seem to lead us into disaster.

In our paper published in Nature last week, we explored this puzzle, using evolutionary game theory to explore the performance of alternative strategies in competition with each other. The results were quite striking. If there were no error in assessments, then an unbiased view of the world is indeed the best strategy. However, in the real world, error in assessments of others or of the tasks ahead of us are commonplace, and this leads to an interesting outcome: As long as the prize at stake in some task is relatively large compared with the costs of attempting to gain that prize, then a bias toward overconfidence turns out to be the winning strategy—doing better than accuracy or excessive caution. The reason it does well is because it biases us to claim and compete for prizes that we might otherwise shy away from.

The model suggests that, if the rewards were large compared with the costs during human evolutionary history, natural selection may have evolved a bias toward overconfidence as an adaptive strategy. But the model also suggests that, whether or not it evolved as part of human nature, wherever benefits are large and costs are small, overconfidence will succeed and spread if people copy or imitate the successful strategy. This can occur very rapidly as a process of cultural evolution in groups, organizations, or societies.

Having been working on a parallel project on evolutionary theories of religion, there seem to be two interesting points of crossover.

First, in general terms, religious “faith” appears to be at least in part about believing in yourself and your abilities. And where that belief is weakened, people often turn to supernatural agents to help find strength. In some situations, any cognitive mechanism that helps to promote faith in your abilities and what you can achieve may be similar to the winning strategy of overconfidence. Overconfidence may not be quite the right word in this context, but the evolutionary logic holds.

Second, one evolutionary theory of religion suggests that a belief in supernatural punishment may have been favored by natural selection because it helps to keep selfish desires in check (the possibility of the watchful gaze of an omnipresent agent makes us think twice about cheating on our fellow man or doing something selfish, even if we think we are alone). However, such a costly belief would only be selected for by evolution if selfish human desires incurred costs that were even greater. In a recent article on supernatural punishment theories in the journal Religion, Brain & Behavior, Jeff Schloss and Michael Murray suggested exactly this. Supernatural punishment may have been useful if humans were overconfident, they point out, about the likelihood of getting away with selfish behavior and cheating. Our new paper adds weight to this idea. Perhaps evolution could not easily suppress overconfidence, either because it was already a deep seated aspect of human nature, or because it remains essential in many other walks of life. Religious beliefs may have offered an important behavioral corrective.

There is a lot of speculation here, but it certainly raises interesting questions for future research. Whether positive illusions are adaptive, and the consequences of this possibility for our picture of human nature, would seem to have some important implications for a wide range of theories of religion.

Dominic Johnson is a reader in politics and international relations at Edinburgh University.

Category: Q&A


3 Responses

  1. Theophile says:

    Hi Dominic,
    You said, “There is a lot of speculation here, but it certainly raises interesting questions for future research.”
    Isn’t that the mantra of evolutionary science?

  2. PeterKinnon says:

    Certainly religion, together with many other aspects of human behavior (theater, sport, graphic arts, for instance) can properly be interpreted as “spinoff” from the overall evolutionary process.

    However, I would argue that technology rather than “culture” should be considered the primary evolving entity.

    This is the model outlined in “The Goldilocks Effect: What Has Serendipity Ever Done For Us?” (free download in e-book formats from the “Unusual Perspectives” website)

  3. Jeff Schloss says:

    Theophile’s comment is a fair one: speculation + future research is, in a sense, the mantra of evolutionary science. In fact, it is the mantra of all science, which generates hypothesis – many quite speculative – that then suggest research that can distinguish between alternative hypotheses or discover new patterns that follow from the proposal. If an idea is not fruitful in this way, it’s not good science. But the way to find out it to launch the idea…

Leave a Reply