Strategies
Mean reversion strategy for stocks: how to build one
By Koen van Goch · Updated
A mean reversion strategy for stocks buys shares after a short, sharp drop and sells them after a bounce, usually within days. On the short side, it does the opposite after a spike. It can get paid for providing liquidity to people who are selling in a hurry. It typically wins often but loses big on the trades that don't bounce, so the exits and sizing matter a lot.
Why can mean reversion get paid?
When investors sell in a panic, for example on a bad headline, a forced liquidation or a sharp market dip, prices can overshoot. A buyer who steps in when others must sell is providing liquidity and can be paid for it when prices settle. The edge rests on that behaviour, not on any indicator.
Typical building blocks of a mean reversion system
- Universe: liquid stocks, often only those in a longer-term uptrend for the long side.
- Setup: a short-term oversold condition, such as several down days or a short-term oscillator at an extreme.
- Entry: often a limit order below the last close, buying into weakness.
- Exit: a bounce condition or a time stop after a few days.
- Sizing: risk-based or volatility-based, with a cap on how many positions open at once. See position sizing in trading.
Risks to respect
Mean reversion has a high win rate and fat-tailed losses: a stock that keeps falling can wipe out many small wins. This is called left-tail risk: most results sit just above zero, but the distribution has a long tail on the losing side. Trend following is the mirror image, with many small losses and a long right tail of big winners.
Clusters of positions opened during a market-wide sell-off are correlated, so several of those tail losses can arrive at once. Results are also sensitive to costs and fills, so test with realistic assumptions.
Trend following
Many small losses, a few big winners
Mean reversion
Many small wins, a few big losses
Illustration of the typical shape, not real data. Each bar shows how often a trade ends with that result.
Why pair mean reversion with trend following?
Trend following tends to do well in persistent moves and struggle in choppy markets. Short-term mean reversion often does the opposite: it's the chop specialist. Run together, with sizes you've tested, they can smooth each other out. My own live portfolio combines trend following, mean reversion and hedge systems for exactly this reason. You can follow it on the live track record.