Claim
Intuition is useful, but if it stays implicit you never get feedback on when it fails. Force every important intuition into an explicit, falsifiable prediction with a number and a horizon, then check it.
Mechanism
Implicit intuition is unfalsifiable: any outcome can be retrofitted into "I knew that." Explicit predictions are anchors. They force the intuition into a shape that can be wrong, which is the only shape that can teach. Over time, calibration emerges, the operator learns where their gut is sharp and where it is biased. Without the explicit step, both gains and losses go unrecorded.
Conditions
Holds when:
- The intuition is about something measurable in a reasonable time.
- The operator is willing to be wrong publicly enough to learn.
- The team treats wrong predictions as data, not failure.
Fails when:
- The variable is genuinely unmeasurable.
- The org culture punishes wrong predictions, so people stop making them.
- The horizon is too long to close the loop within decision relevance.
Evidence
"I think this positioning will resonate" is implicit. Explicit: "We expect 15% CTR on this LP variant, and we'll know we're wrong if it's <8%."
Annie's framing: explicitness forces specificity, specificity enables falsification.
· Annie Duke on Lenny's Podcast, 2026-04-28
Signals
- Decision documents include explicit predictions with numbers and horizons.
- Post-mortems compare actual outcomes to recorded predictions, not to vague memories.
- Operators develop measurable calibration over months, better tracking, fewer surprised reactions.
Counter-evidence
Forcing every intuition into a number can produce false precision and crowd out genuine ambiguity-tolerance. Some operators reason better by analogy and metaphor than by quantified prediction.
Cross-references
- Pre-mortems only work if you commit kill criteria before starting, the launch-gate companion
- There is no such thing as a long feedback loop, find a correlated short signal, how to keep the horizon short