Recency Bias
Overweighting recent outcomes when judging what is likely to happen next.
Also called: recent bias
Written by Javier Sánchez Ros
In plain language
A few recent wins make a strategy feel better than the data supports; a few recent losses make a sound strategy feel broken.
It drives the most damaging cycle in trading: sizing up after a good run and abandoning the approach during an ordinary drawdown.
Normal variance is far larger than intuition suggests. Runs of five or six consecutive losses are entirely expected at typical win rates.
Worked through
Sizing up after six wins, quitting after six losses
- After a 6-win streak
- risk raised to 3%
- What the long-run edge was
- unchanged
- After a 6-loss streak
- strategy abandoned
- What the long-run edge was
- unchanged
Six in a row in either direction is entirely ordinary for a strategy that wins about half the time — over a few hundred trades it will happen repeatedly. It carries no information about the method, and it feels like it carries all of it.
So the largest position gets taken at the point where nothing has been learned except that variance clustered favourably, and the method gets discarded at the point where variance clustered the other way. Both decisions are made on the same non-evidence, and they are timed to do the maximum damage.
The asymmetry is what makes it costly. Sizing up before a reversion means the losses that follow are taken at triple weight; quitting before a recovery means the winning stretch that follows is not participated in at all.
The defence is to fix the sample size in advance. Judge a strategy on a hundred trades and size on its long-run expectancy, not on the last six — and hold both numbers in writing, where a good streak cannot quietly revise them.
Why it matters
Recency bias causes traders to make their largest bets right before mean reversion and to quit right before recovery.
Common mistakes
- Increasing risk after a winning streak.
- Abandoning a strategy after a normal-length losing run.
- Judging performance over ten trades rather than a hundred.
Keep exploring
These concepts are connected. Understanding one usually makes the next one easier.
The average amount you expect to win or lose per trade over a large sample.
The percentage of your trades that close at a profit.
The decline from an account’s peak value to its lowest point before a new peak.
A record of every trade, including the reasoning behind it and the result.
The fixed share of your account you are willing to lose on any single trade.