Ideas
Intuition — trust your gut
Your gut has a good track record in some situations and a poor one in others. The skill is knowing which kind of situation you are in.
"Trust your gut" is advice that is given very freely and should be taken with significant care. Not because intuition is unreliable, but because it is selectively reliable in ways that matter a great deal.
Intuition is fast pattern recognition operating below conscious awareness. When the patterns it is drawing on are accurate representations of the current situation, the output is useful. When the patterns are based on incomplete data, biased experience, or a situation that is different from the ones the intuition was calibrated on, the output is a confident wrong answer.
When intuition tends to be reliable
Research on expert intuition — from Klein's work on naturalistic decision-making to Kahneman's summary in Thinking, Fast and Slow — converges on two conditions that make intuitive judgment reliable.
High-validity environments: Situations where there is a reliable relationship between cues and outcomes, so that the environment provides accurate feedback. Chess is the clearest example. Medical diagnosis in pattern-based specialties is another. The gut-level sense that something is wrong with a patient, developed over thousands of similar cases, is a real signal.
Sufficient practice with feedback: The pattern library that powers intuition is built from experience. But experience only builds good pattern libraries if it includes feedback — if you actually found out whether your assessment was correct. Many domains appear to offer this feedback but do not, because the feedback is delayed, ambiguous, or filtered in ways that prevent accurate learning.
Where both conditions apply, intuition deserves significant weight. Where one or both are absent, it deserves scrutiny.
When intuition tends to mislead
Low-validity environments: Situations where the relationship between cues and outcomes is weak, noisy, or systematically misleading. Long-term predictions in complex social or economic systems are a good example. The gut feeling about where the stock market is headed is not a reliable signal because the environment does not provide the kind of consistent feedback that would build accurate intuition.
Novel situations: Intuition is calibrated on past experience. In situations that are genuinely new — a technology category that did not exist before, a social situation with no precedent — the patterns the intuition draws on may not apply.
High-stakes, low-frequency decisions: Decisions you make very rarely provide no opportunity to build an accurate intuitive sense. The gut feeling about a major life decision — a career change, a significant relationship choice — is not backed by the same pattern library as expert judgment in a well-practised domain.
Motivated reasoning: Intuition is vulnerable to desires. The confident feeling that something is a good idea may be the intuition or it may be the thing you want to be true, experienced as an intuition. These are functionally indistinguishable from the inside.
A practical approach
Rather than a blanket instruction to trust or distrust intuition, the more useful approach is to ask a few specific questions when a gut feeling is prominent.
What is the validity of this environment? Is there a reliable relationship between the cues I am responding to and the outcomes that matter? Do I actually get feedback on whether my judgments in this domain are correct?
How much relevant experience do I have? Have I made similar judgments many times, and have I found out whether I was right?
What would I think if the feeling pointed the other way? If the gut feeling were pointing toward the other option, would I take it equally seriously? If not, the feeling may be confirming what you already want rather than providing independent information.
Is there cheap analysis available? In some cases, a simple check — a quick calculation, a second opinion, a day's delay — provides enough additional signal to either support or challenge the intuition at low cost.
None of this means that intuition should be overridden by analysis as a default. In domains where your intuition has been well calibrated, the gut feeling often outperforms deliberate analysis, particularly under time pressure or information overload. The point is that "trust your gut" is a conditional instruction, not a universal one.