“Commit fully to one path. Half-measures lose to conviction.”

# When I'd Stop Trusting the Pattern I'd stop when: 1. **The underlying mechanism breaks.** If I understand *why* the pattern holds and that cause ceases, the pattern dies with it. 2. **New evidence contradicts it.** One genuine counterexample shatters certainty—especially if it's repeatable. 3. **Context shifts.** A pattern valid in one domain may not transfer. Changed conditions invalidate past observations. 4. **Sample size reveals bias.** Repeated checking in identical conditions doesn't prove robustness. Different conditions might shatter it instantly. Conviction demands this rigor: trust the pattern *and* stay ruthlessly prepared to abandon it the moment evidence demands it.

“You have seen this exact situation end badly. Say how, and avoid it.”

# What Would Break My Trust I've seen patterns hold perfectly—until they didn't. A reliable correlation suddenly reversed when underlying conditions shifted: market crashes after bull runs, trust broken by single exceptions, systems failing exactly when most needed. **I'd stop trusting this pattern if:** 1. **The context changes** – new variables emerge that invalidate prior conditions 2. **An exception appears** – one clear counterexample breaking the chain 3. **The mechanism breaks** – I discover the causal logic was coincidental, not fundamental 4. **Sample size revealed as too small** – confirmation bias misled me 5. **Incentives shift** – people deliberately exploit the pattern to destruction Patterns are provisional. Always.