中文

Cognitive biases · Judgement

Overconfidence

We often think we know more, or do better, than we actually do.

What it is

Overconfidence is a gap between how good we think our judgment or performance is and how good it actually is. Moore and Healy (2008) point out that the word covers three different things:

  • Overestimation: thinking you did better than you did.
  • Overplacement: thinking you are better than others.
  • Overprecision: being too sure that your beliefs or estimates are correct.

An everyday example

Chris is asked how long it will take to drive to a friend’s wedding. “Two hours, give or take ten minutes,” Chris says confidently. The trip takes two hours and forty minutes because of road work and a missed exit.

The estimate itself was not wild. The problem was the narrow range: Chris left almost no room for the many small things that can go wrong.

Why it happens

We know our own intentions and effort in detail, but we often lack good feedback on how accurate we were in the past. We also tend to underestimate how much we do not know, so our ranges come out too narrow.

Moore and Healy (2008) argue that some patterns follow from imperfect information. On hard tasks, people tend to overestimate their own scores but believe they did worse than others. On easy tasks, the pattern reverses. Overconfidence is not simply “everyone thinks they’re great”.

How to counter it

  • Widen your ranges. When estimating, give a low and a high figure you would be genuinely surprised to fall outside of.
  • Keep a track record. Write down predictions and check them later.
  • Imagine it went wrong. Before committing, list the most likely reasons a plan could fail.
  • Ask for outside views. Compare your estimate with how long, or how well, similar efforts went for others.

What the research shows

In a classic study, most American and Swedish participants rated themselves as safer and more skillful than the average driver in their group (Svenson, 1981). A large preregistered replication with over 1,200 American participants found the same pattern (Koppel et al., 2023).

Moore and Healy’s (2008) review found that overprecision is the most persistent of the three forms. Overestimation and overplacement depend more on the task: they are common on hard or easy tasks respectively, and can reverse.

Confidence is not the enemy. It helps people act. The aim is calibration: confidence that matches how often you turn out to be right.

Related biases

  • Planning fallacy — We expect our own tasks to go faster and smoother than similar tasks usually do.
  • Illusion of control — We often feel we can steer outcomes that are mostly down to luck, especially when we get to make choices.
  • Confirmation bias — We tend to look for, and believe, evidence that fits what we already think.

Sources

  1. Moore, D. A., & Healy, P. J. (2008). The trouble with overconfidence. Psychological Review, 115(2), 502–517. https://doi.org/10.1037/0033-295X.115.2.502
  2. Svenson, O. (1981). Are we all less risky and more skillful than our fellow drivers? Acta Psychologica, 47(2), 143–148. https://doi.org/10.1016/0001-6918(81)90005-6
  3. Koppel, L., Andersson, D., Tinghög, G., Västfjäll, D., & Feldman, G. (2023). We are all less risky and more skillful than our fellow drivers: Successful replication and extension of Svenson (1981). Meta-Psychology, 7. https://doi.org/10.15626/MP.2021.2932

All biases