Knowing something without being able to say how you know it — reliable under two specific conditions, and unreliable outside them.
// Where it sits
Chess and firefighting sit at the left end. Long-horizon forecasting sits at the right.
What it really means
Two researchers who had spent careers disagreeing about intuition sat down to write a paper about it. Daniel Kahneman thought expert hunches were mostly overrated. Gary Klein had spent years documenting fireground commanders making excellent split-second calls they could not explain.
The paper is titled A Failure to Disagree.
What they found is the most useful thing anyone has said about the subject, because it is not a verdict — it is a filter. Intuitive expertise is trustworthy when two conditions both hold:
One — the environment is regular enough that valid cues exist. There has to be a real pattern in the world to learn. Chess positions have one. Fire behaviour has one. Long-horizon political or market forecasting largely does not.
Two — you have had prolonged practice with rapid, unambiguous feedback. Not just exposure. Corrected exposure. You have to find out you were wrong, quickly and clearly, thousands of times.
Both, or neither. An expert in a low-validity environment is not a weak intuiter — they are a confident one with nothing to have learned.
Which raises the obvious question: can I feel the difference?
No. That is the finding people least want. Subjective confidence is an unreliable indicator of whether an intuitive judgement is valid. Leach and Weick tested it directly: people who believed strongly in their own intuition performed no better on implicit-learning tasks than people who didn't. The strength of the hunch and the quality of the hunch are separately distributed.
So the practical question is never how sure do I feel? It is: is this a domain where the feeling could have learned anything?
Where it comes from
The root is visual — intuition is literally looking at something rather than reasoning about it. The word was theological for centuries before it was psychological.
Myths & misconceptions
Trust your gut — expert intuition is a mysterious faculty that either works or doesn't.
Kahneman and Klein spent years disagreeing and then published where they converged. Intuitive expertise is trustworthy when, and only when, two conditions both hold: the environment is regular enough that valid cues exist, and the person has had prolonged practice learning those regularities with rapid, unambiguous feedback. That is not a hedge — it is a testable filter you can apply to any domain before trusting anyone's hunch, including your own.
You can tell when your own intuition is trustworthy — strong hunches are better hunches.
Subjective confidence was identified as an unreliable indicator of the validity of intuitive judgements. Leach and Weick tested it directly and found that participants who believed their intuitions were strong performed no better on implicit-learning tasks than those with less faith in intuition. The feeling of a good hunch and a good hunch are separately distributed.
Compare & contrast
Intuition vs instinct
Instinct is unlearned; intuition is over-learned. That is exactly why the two conditions matter — intuition is compressed experience, so it inherits the quality of the environment the experience came from. Bad environment, confident nonsense.
| // | Intuition | Instinct |
|---|---|---|
| Learned | Yes, heavily | No |
| Domain-specific | Yes | No |
| Transferable | Poorly | n/a |
| Improves with feedback | Yes | No |
How it connects
- Overconfidence Effectwhy the strength of a hunch tells you so little about its accuracy.
- Hindsight Biasthe mechanism that keeps a bad hunch feeling like it was right all along.
Tell it apart
Questions people ask
- When should I trust my intuition?
- When the situation is regular enough to contain valid cues, and you have practised in it long enough with fast, clear feedback to have learned them. Chess and firefighting fit. Long-horizon prediction in noisy environments does not — and no amount of experience there converts into reliable intuition, because the feedback never arrives clean.
- Does feeling certain mean I'm right?
- No. Subjective confidence was explicitly identified as an unreliable indicator, and a 2018 study found belief in one's own intuition did not predict actual intuitive accuracy.
- What is an adversarial collaboration?
- Two researchers who publicly disagree agreeing in advance to work jointly and publish the result, including where they still disagree. The 2009 paper is the field's showcase example — and its title says the quiet part: A Failure to Disagree.