Cognitive bias
A systematic error in thinking: the mind departs from the accurate answer not at random but in a predictable direction, and does it the same way every time.
The load-bearing word is "systematic". Random errors cancel out: half your estimates run high, half run low, the average lands near the truth. A bias does not behave that way — it pulls everyone in the same direction every time, which means the error can be predicted before anyone makes it.
Amos Tversky and Daniel Kahneman named the effect in 1974. Their point was not that people are foolish but the opposite: bias is the price of speed. The mind judges by simplified rules that, in ordinary life, return a good answer instantly. The failure comes when the situation is unusual and the rule fires anyway.
The uncomfortable part: knowing about a bias barely protects you from it. People confidently spot bias in others and miss it in themselves — the trait is well documented and has its own name, the bias blind spot. What helps is not effort but procedure: decision criteria written down before the options appear, an outside opinion, a day between question and answer.
The five you will meet most often
- Confirmation — you look for what agrees with the decision you already made and walk past the rest.
- Anchoring — the first number mentioned drags every later estimate toward it, even a number pulled from thin air.
- Availability — an event feels likely if it comes to mind easily, which is why plane crashes frighten people more than the statistics warrant.
- Sunk cost — you keep investing in the lost thing precisely because you already invested.
- Hindsight — once the outcome is known it looks obvious, though nobody called it in advance.
Commonly confused with
With a logical fallacy. A fallacy is a defect inside the reasoning: it sits in the text of the argument and you can point at it. A bias acts earlier, before any argument — it decides what you noticed and retained in the first place. So the reasoning can be flawless and the conclusion still wrong, because the mind picked the inputs for you.
Common questions
There is no agreed list. Popular charts show around two hundred, but that is a collection of labels from unrelated papers rather than a taxonomy, and many entries describe the same effect under different names. Five or six that actually touch your decisions are worth more than two hundred you can recite.
Remove them, no; route around them, yes — and not by willpower. External scaffolding works: criteria written down before the options exist, a person you owe an explanation to, a rule against answering on the day the question arrives.
Usually not. A fast call on incomplete data is exactly what they are for, and in everyday situations the call is fine. The systematic error surfaces where the case is rare, the stakes are high and feedback arrives late — which is to say, in the decisions that matter.
