Mean Reversion

Mechanism

The EA measures how far price has deviated from a moving average or Bollinger Band centre. When deviation exceeds a threshold (typically 1.5–2.5 standard deviations) confirmed by a momentum oscillator (RSI or Stochastic in extreme territory), an entry fires in the opposite direction. The take-profit is the mean itself; the stop is placed beyond the extreme that would invalidate the reversion thesis.

Suitability

Best on pairs with documented mean-reversion profiles: EUR/CHF, USD/CHF, EUR/GBP. Structurally unsuited to momentum currencies during trend phases (EUR/USD in 2022, USD/JPY in 2023). An ADX filter below 25 is almost always necessary — without it, mean-reversion entries in trending markets produce deep, slow-recovering drawdowns. Lower return than trend strategies but higher win rate (65-75%) and lower drawdown.

Notes

Mean reversion bets on elastic, not momentum: that a price stretched far from its average is more likely to snap back than to keep stretching. A mean-reversion EA waits for a statistically significant deviation — typically 1.5 to 2.5 standard deviations from a moving average or Bollinger centre — confirms it with an oscillator in extreme territory, then enters against the move, targeting the mean itself.

It is the natural opposite of trend-following, and its strengths and weaknesses are mirror images. Win rate is high (often 65-75%) because most stretches do revert, and individual trades are short. But the losses, when they come, are the dangerous kind: a market that stops reverting and starts trending runs straight through the entry, and counter-trend positions bleed slowly. An ADX filter below 25 is almost always required to keep the EA out of trends.

When evaluating a mean-reversion EA on this catalogue:

1. Pair selection — it belongs on structurally range-bound pairs (EUR/CHF, USD/CHF, EUR/GBP), not momentum majors.
2. Trend filter — is there an ADX or regime filter to suppress entries in trends?
3. Win rate vs worst loss — a high win rate means little if one trending month erases a quarter of gains.
4. Stop discipline — the stop beyond the extreme is what separates mean reversion from an accidental martingale.