More Than You Know

5 ideas

1. Judge decisions by process, not outcome

   In probabilistic fields a good decision can produce a bad result and a bad decision a good one, so outcomes are a noisy grade of skill. Crossing process quality with outcome quality gives four cells: good process and good outcome is deserved success, good process and bad outcome is a bad break, bad process and good outcome is dumb luck, and bad process and bad outcome is poetic justice. Evaluators who reward only outcomes end up reinforcing dumb luck and punishing bad breaks.

2. Expected value beats frequency of being right

   What matters is not how often you are right but how much you gain when right compared with how much you lose when wrong. Expected value is the probability of each outcome multiplied by its payoff, summed across outcomes. A strategy that is right only 30% of the time can dominate one that is right 70% of the time if its wins are large and its losses small, yet people instinctively anchor on hit rate.

3. Crowds are wise only while diverse

   A crowd's collective estimate beats most individuals when three conditions hold: participants' errors are diverse and independent, there is a mechanism to aggregate their views, and there are incentives to be right. Markets are efficient when these conditions hold and inefficient when they fail. Booms and crashes occur when imitation and herding collapse diversity, so individual errors stop cancelling and instead compound in one direction.

4. Treynor's jelly bean jar experiment

   The average of the group's guesses was 871, and only one student guessed closer than the group average. The case shows how the independent errors of many people cancel out, and it underpins the argument that market prices can aggregate information better than almost any single participant.

5. Markets as complex adaptive systems

   A stock market works like an ant colony: many heterogeneous agents following local rules produce aggregate behaviour that none of them intends and that cannot be deduced from studying any single agent. Because the whole emerges from the interactions, cause and effect are nonlinear and hard to trace. Tidy explanations of why the market moved on a given day are therefore usually after-the-fact stories rather than real causes.

Save and mark ideas in the app