↩️ Mean Reversion

Mean reversion is the counterpunch to trend-following. Markets overreact: a burst of buying or selling stretches price far from its recent average, and more often than not it snaps back. This strategy quantifies 'too far' with Bollinger Bands — a moving average with bands drawn a set number of standard deviations away — and fades the stretch, targeting the return to the middle.

It's the natural opposite of the Trend Breakout scanner: one buys strength, the other sells it. Running both is a classic way to hold edges for different market regimes.

How the strategy works

1. On 1-hour bars, draw Bollinger Bands: a moving average ± a multiple of standard deviation.
2. When price closes beyond a band, it's statistically over-stretched.
3. Enter in the opposite direction, targeting the middle band — the mean.
4. The stop is a multiple of ATR beyond the entry.
5. By default a trend filter only allows fades in the direction of the bigger trend — fighting a strong trend is how mean reversion dies. (It can be disabled to fade both ways.)

When it shines — and when it doesn't

Mean reversion thrives in range-bound, choppy markets — exactly where breakout strategies bleed. Its weakness is the strong trend: what looks 'over-stretched' in a real bull run keeps stretching, which is why the trend filter matters more than any other setting.

Key terms

Bollinger Bands

A moving average with bands drawn ± k standard deviations around it — a live map of 'normal' vs 'stretched'.

The mean

The middle band (the moving average) — the magnet this strategy trades back toward.

Fading

Trading against the recent move — selling a spike up, buying a spike down.

σ (standard deviation)

The statistical unit of 'how unusual is this move' that sets the band width.

How to judge it

Mean reversion wins often but small — the opposite profile to breakouts. A high win rate alone can hide an account-killer, so judge win rate and profit factor together, and watch max drawdown in the backtest: the losses cluster when a range breaks into a trend.