Leaderboards select for luck
Posted: Fri Oct 30, 2026 12:52 pm
Leaderboards select for luck
Trading forums and signal platforms rank traders by recent performance. The top ten are visible, admired, and copied. What the list hides is the thousands who started and vanished.
Consider a thousand traders flipping coins with a mild edge. After a month, a few dozen will have remarkable streaks by chance alone. Rank them, and the leaderboard will be full of apparent geniuses who are just the lucky tail of a large distribution.
Survivorship bias compounds the issue. Accounts that blow up disappear, and the list shows only survivors. The average performance of everyone who tried is far below that of those on display.
Better evidence: a long record across different conditions, verified by statements, with drawdowns shown honestly.
A healthy default is to treat any impressive curve as a hypothesis, not a fact, until it has survived a regime change.
What would a track record need to show before it earned real trust?
Trading forums and signal platforms rank traders by recent performance. The top ten are visible, admired, and copied. What the list hides is the thousands who started and vanished.
Consider a thousand traders flipping coins with a mild edge. After a month, a few dozen will have remarkable streaks by chance alone. Rank them, and the leaderboard will be full of apparent geniuses who are just the lucky tail of a large distribution.
Survivorship bias compounds the issue. Accounts that blow up disappear, and the list shows only survivors. The average performance of everyone who tried is far below that of those on display.
Better evidence: a long record across different conditions, verified by statements, with drawdowns shown honestly.
A healthy default is to treat any impressive curve as a hypothesis, not a fact, until it has survived a regime change.
What would a track record need to show before it earned real trust?