Why confidence tends to run ahead of what the evidence supports

Feeling sure and being right are separate states, and the first is much easier to arrange than the second.

There is a gap between how certain something feels and how much support it has, and the gap is not random: it widens with familiarity. Something you have thought about often feels settled, because thinking about it repeatedly is the same experience as thinking about it thoroughly, from the inside. You have driven the same route for six years, so you know it is the fastest. Do you. You know it is fast enough that you stopped checking, which is a different claim, and the confidence you have does not distinguish between them.

A modest correction is to attach a rough number to the feeling. Not a false precision, just a band: am I as sure of this as I am that the shops shut at five, or as sure as I am that it will rain on Thursday. Putting a belief on that crude scale forces a comparison, and comparisons are where overconfidence starts to look silly. The other useful question is: how would I know if I were wrong about this. If no answer comes, the confidence is not resting on evidence at all. It is resting on the fact that nothing has contradicted it lately, which is a much thinner floor than it feels like.

It is worth saying that under confidence is a real failure too, and this magazine has no interest in a culture where nobody will commit to anything. Endless hedging is not modesty, it is a way of never being checkable. The aim is calibration: being sure in proportion to your grounds, and saying so. Someone who says they are fairly confident and is right about four times in five is more useful than someone who is certain about everything and right about the same proportion, because you can do something with the first person's estimates.