Search deliberately for the black swan

Look for the single example that would overturn your generalization.

Why it works

Popper pointed out that no number of confirming instances proves a universal claim — one black swan refutes "all swans are white." Searching for disconfirming examples rather than confirming ones is a more efficient path to testing a claim: confirming examples accumulate indefinitely without adding certainty, while a single counterexample can settle the matter. This asymmetry means the smart search strategy is to look for what would break the rule, not what fits it.

How to do it

1. State the belief as a universal claim: "All X are Y" or "X always leads to Y."
2. Search specifically for cases where X occurred but Y did not.
3. If you find one, update: either the claim is false, or you need to narrow its scope.
4. If you genuinely cannot find one, assess whether that is because the claim is strong or because you searched selectively.

Evidence

Wason’s selection task (1968) is a classic demonstration of how people naturally prefer confirming over disconfirming tests; training on falsification search improves performance on logical reasoning tasks. (observational)

Abstract logical training does not reliably transfer to content-rich real-world reasoning; the black-swan search must be applied deliberately in each domain.

Sources

• Wason (1968), "Reasoning about a rule," Quarterly Journal of Experimental Psychology — original selection task

Common mistake

Searching in the set of cases you already know well, where confirming instances are overrepresented because of how you built your knowledge base in the first place.