judgement
Keep score
The rule
You only get better at reading the market by grading your own dated, numbered calls. Like keeping a scorebook, and most people never do.
Where it flips
But scoring can mislead over tiny counts. Ten calls is a hint, not a verdict. The fix is to keep scoring across many calls and watch the trend, not one streak.
Write each view as a percentage with a deadline, then check it later. When your '80% sure' calls come true only 60% of the time, you are too confident. And that overconfidence, not bad luck, is what quietly makes people bet too big.
A worked example
A composite investor writes down ten '80% confident' calls over a year; six come true. The honest lesson is to size to the 60% you deliver, not the 80% you feel. [illustrative]
How to spot it
- ·no written record of past calls
- ·confidence stated without a number or a date
- ·only the wins are remembered
Philip Tetlock · Superforecasting