judgement
The wrong-side-of-maybe fallacy
The rule
A 70% guess is not "wrong" when the 30% happens. Judge odds across many cases, not one single result.
Where it flips
This idea can become a shield: "I said only 70%" to excuse every miss. The fix is to still check your hit-rate over many calls, so a truly bad guesser cannot hide behind odds.
People call a guesser wrong the moment the less-likely thing happens, as if 70% meant sure. But a 70% call openly said the other thing could happen three times in ten. The only honest test is the long record. Across many 70% calls, did about 70 in 100 come true? Judging one such call by its single result is a common mistake.
A worked example
A guesser said 75% the market recovers. It didn't, and the crowd calls her a fool. But across all her 75% calls, three in four came true. She was well-calibrated, and this was the expected one-in-four. [illustrative]
How to spot it
- ·a single outcome used to judge a probability call
- ·"70%" treated as if it meant certain
- ·no look at the long record
Philip Tetlock & Dan Gardner · Superforecasting