Mean Reversion Strategy — Trading the Snap-Back from Extremes

Mean reversion is the trading-strategy expression of a statistical observation: price oscillations frequently reverse from extreme readings back toward a moving average. RSI readings above 70 or below 30, prices touching Bollinger Band extremes, or significant deviations from a long-period moving average all signal 'price has moved too far in one direction' and create potential entry signals in the opposite direction.

The strategy's logic: extreme readings imply imbalanced buying or selling pressure that's likely to exhaust and reverse. Buying when RSI is below 30 effectively means 'sellers have been dominant; eventually buyers will return and price will recover'. The strategy bets on the timing — entering at the extreme and exiting at the moderate level.

Mean reversion is the dominant winning strategy class in ranging markets. When price is oscillating within a defined range, every approach to the range boundary creates a mean-reversion opportunity. The same strategy is the dominant losing approach in trending markets, where 'overbought' just means 'trend continuing' and 'reverting to mean' means 'getting stopped out on continued trend'.

Entry triggers commonly combine: (1) primary oscillator signal (RSI, Stochastic, CCI at extreme), (2) structural confirmation (price at support/resistance level, Bollinger Band touch), (3) reversal candle pattern (engulfing, pin bar against the move), (4) regime filter (don't trade if ADX > 25 indicating trending market).

Exit logic: most mean-reversion strategies use fixed take-profit at the middle Bollinger Band (20-SMA typically) or the 50-RSI level. Trailing stops aren't typical because the strategy targets quick reversion rather than capturing extended moves. Stops at 1.5-2× ATR beyond entry give the position room for the extreme to extend modestly before invalidation.

The regime filter is critical and often missing in retail mean-reversion EAs. Trading mean reversion during a strong trend produces consecutive losing trades — RSI in oversold reads in a downtrend keeps reading oversold as price continues lower. Adding an ADX or trend-direction filter that prevents trading during trends improves win rate substantially (from ~55% no-filter to ~70%+ with proper filter).

Position sizing: fixed-fractional at 1-2% per trade. Mean-reversion's higher win rate means tolerance for slightly larger per-trade risk than trend-following — the win-rate margin absorbs more variance. But this doesn't justify careless sizing; the consecutive losses during regime shifts can still erase months of gains in days.

Mean reversion as systematic strategy dates to 1950s-1960s statistical arbitrage research. Edward Thorp's work in the 1960s-1970s formalised mean-reversion principles applied to financial markets. The strategy class has been studied extensively in academic literature, with consistent findings: mean reversion has positive expected returns in markets exhibiting auto-negative-correlation (range-bound behaviour) and negative expected returns in markets with auto-positive-correlation (trending behaviour).

In forex specifically, mean-reversion strategies were dominant retail trading approaches in the 1990s-2000s when major pairs were generally range-bound around fundamental fair values. The 2008 financial crisis and subsequent quantitative easing periods shifted forex toward more trending behaviour, making mean-reversion-only strategies underperform for extended periods.

Current state (2020-2026): mean-reversion still works in clearly ranging regimes (e.g. consolidation periods between major macro events) but the regime-identification challenge is harder than in earlier eras. Many traders run mean-reversion only as part of a diversified portfolio with trend-followers, sized inversely so that whichever regime is active drives that period's returns.