Mean Reversion

Mechanism

The EA measures how far price has deviated from a moving average or Bollinger centre, usually in standard deviations or ATR units, and confirms the stretch with an oscillator in extreme territory. Entry is against the move; the target is the mean, and the stop sits beyond the level that would invalidate the reversion thesis. The distance between those two exits fixes the shape of the record. A target close to the entry produces a high win rate and a payoff ratio well below 1; a target at the full mean produces the reverse. Neither setting creates an edge the entry does not have.

Suitability

Best where the symbol spends most of its time inside a range and returns to a definable centre — historically the CHF crosses, EUR/GBP and AUD/CAD. Structurally unsuited to momentum currencies during trend phases, when the stretch keeps stretching and a counter-trend position is held against a move that does not return. A trend or regime filter is close to mandatory, and the direction of the higher-timeframe trend matters more than the oscillator level. Suitable for a trader who can accept long losing runs that arrive in a block, and who reads the stop distance before the win rate.

Notes

Mean reversion bets on elastic rather than momentum: that a price stretched far from its average is more likely to snap back than to keep stretching. The EA waits for a statistically significant deviation — typically 1.5 to 2.5 standard deviations from a moving average or Bollinger centre. It confirms the stretch with an oscillator in extreme territory, then enters against the move and targets the mean itself.

That describes the entry, and the entry is the part most people understand. The description usually bolted onto it — mean-reversion systems win most of their trades and lose big occasionally — is a claim about where the exits sit, not about the mechanism. This page covers how the machine is built inside MetaTrader 5 and what the Builder’s own measured reversal templates show. It ends with how to stress-test one before any capital is exposed.

How it works: the entry is a distance, the exit is the shape

The entry side is a measurement, and the three common constructions are close to interchangeable:

• Deviation from a moving average. Price a set number of ATR units or standard deviations from a mean. The whole strategy in one line.
• Bollinger touch or close beyond the band. The same idea with the deviation calculated for you. A close beyond the band is stricter than a touch, and it changes trade count more than win rate.
• Oscillator extreme with a regime filter. RSI or Stochastic past a threshold, gated by a reading that says the market is not trending. The filter is the difference between a mean-reversion EA and one that fades trends until it is stopped out of all of them.

The exit has two parts: a target at or near the mean, and a stop beyond the level that invalidates the thesis. Expectancy per trade is win rate × average win − (1 − win rate) × average loss, and once the two exit distances are fixed the terms stop being independent. Pull the target close and the win rate rises while the average win shrinks below the average loss. Push it out to the full mean and the reverse happens.

The classic reversion arrangement is a near target and a wide stop — a high win rate paid for with a payoff below 1. The ten reversal templates measured below ship the other way round, with a target twice the stop (30/60 or 40/80 pips). That breaks even at a 33.3% win rate before costs, and it means the stock baseline below measures the entry under a trend-shaped exit.

What the Builder’s reversal templates show at stock settings

The EA Builder prints a measured baseline on the template cards that have one, and ten of its reversal templates do. In our own stock-settings baseline of the Builder’s templates — a single USD/JPY M5 run over one year, default inputs, no optimisation — most finished below a 1.0 profit factor. The reversal family is no exception.

Test conditions
Source	Baseline printed on each template card in the EA Builder
Symbol / timeframe	USD/JPY M5 (Exness MT5); the Dual-Pair basket trades EUR/USD and GBP/USD
Period	2025-06-01 – 2026-06-09
Model	M1 OHLC (bar-level, not real ticks)
Deposit / inputs	10,000, template defaults, no optimisation

Template	Entry	Stop / target (pips)	Profit factor	Trades
MA Pullback Buy	Price 0–20 pips above EMA 50 and RSI below 45, buys only	40 / 80	1.10	159
Envelope Reversal	Bid beyond a 0.1% envelope of SMA 20	30 / 60	0.98	986
Bollinger Bounce	Touch of the 20-period, 2σ band	40 / 80	0.97	632
Stoch + BB	Stochastic below 20 at the lower band (mirrored for sells)	30 / 60	0.96	896
Pin Bar Reversal	Bullish pin bar and RSI below 35, buys only	40 / 80	0.96	179
RSI Reversal	RSI 14 below 30 / above 70	30 / 60	0.94	682
Range Filter Reversal	RSI 30/70 only when the 10-bar range is 10–80 pips	40 / 80	0.90	573
MFI Reversal + StdDev	MFI 20/80 only while StdDev 10 is below StdDev 40	40 / 80	0.80	362
3-Strategy Portfolio (Mean-Reversion Mix)	RSI + Bollinger + Envelope in one EA	as each part	0.92	2,287
Dual-Pair RSI Basket	RSI 30/70 on EUR/USD and GBP/USD, up to four positions	30 / 60	0.87	12,953

Four things follow, and the folklore predicts none of them.

The exit dial moves the shape, not the edge. The Builder records a stop-and-target search for RSI Reversal and Bollinger Bounce on USD/JPY H1. Each was run with twelve combinations — stops of 20 to 50 pips, targets of 20 to 100 — over three windows: 2018–2021, 2022–2024 and January 2025 – June 2026. The best in-sample profit factor of either template was 0.91. That grid includes near targets and far ones. When no placement wins even in the window it was chosen on, the entry has nothing for the exit to shape.

A filter is not automatically an improvement. Range Filter Reversal adds a 10-bar range gate to the RSI 30/70 entry and scores 0.90, against 0.94 for the plain version — though the stop and target differ too, so it is not a clean pair. MFI Reversal + StdDev, which only fades while volatility is contracting, is the weakest single-symbol template in the table at 0.80. A gate that removes trades removes good ones as well as bad ones.

The only template in the table above 1.0 fades in the trend’s direction. MA Pullback Buy buys dips only while price sits above a 50-period EMA, and scores 1.10 over 159 trades. It is a pullback entry — a reversion entry on a trend filter — and it never stands against the longer-term move. It still falls short of the Measure panel’s printed pass bar (profit factor at least 1.2, at least 60 trades, drawdown no worse than 20%).

Stacking reversal methods concentrates the risk. The three-method portfolio scores 0.92 — no better than its parts, because all three fade the same stretch. The Dual-Pair RSI Basket runs one RSI rule on two correlated pairs with up to four positions open. It produced 12,953 trades and a -36.38% drawdown, the deepest of any template in the Builder’s baseline. Correlated reversal positions are one bet placed several times.

Market conditions: when mean reversion pays and when it does not

The bad environment is not a losing market — it is a directional one, the exact mirror of how trend-following fails in a range. The two lose money in different weather. It is also why losing runs cluster. Ten consecutive losses is not ten unrelated mistakes; it is one trending stretch met by the same rule ten times.

What caused the deviation matters as much as its size. A price stretched by thin liquidity is a candidate for reversion; a price stretched by a rate decision has been repriced, and there is no old mean to return to. No indicator separates the two, which is why a news filter matters more here.

The same elastic assumption drives statistical arbitrage, applied to the spread between two instruments instead of one price, and it breaks the same way. On a single pair such as EUR/GBP, the relationship assumed is the pair’s own history with its average.

Parameters and settings in MT5

In the Builder each of these is a block with its own inputs, and every value becomes an input in the compiled EA.

Block and input	Typical starting point	What it controls
Bollinger Bands / Envelope period, deviation	Long enough that the mean is a level, not a lagging copy of price	What the EA treats as the value price should return to
RSI / Stochastic / MFI thresholds	Confirmation only, not the trigger on its own	Filters stretches that are still accelerating
Market Regime Filter Allowed regime = Range	ADX threshold around 25, tested rather than assumed	Suppresses entries in trends — the single most important input on this page
MA Distance Filter Price side, Max pips (0=off)	Fade only on the side of a longer MA	Turns a counter-trend fade into a pullback entry
Open Buy / Open Sell TP (pips)	At the mean, or a fraction of the distance to it	Together with the stop, sets win rate and payoff ratio
Stop Loss SL Mode	Beyond the invalidation level, or an ATR multiple	The line between mean reversion and an accidental martingale
Spread Filter Max Spread (points)	Just above the symbol’s normal spread	Extremes and wide spreads arrive together; this blocks the worst fills
Position Filter Max Positions	1 per symbol	Stops the rule from stacking the same fade

Three MT5-specific realities decide whether a promising backtest survives contact:

• The stop has a minimum distance, and this strategy wants a wide one anyway. The broker rejects a stop closer than SYMBOL_TRADE_STOPS_LEVEL. The EA the Builder generates checks that level, and SYMBOL_TRADE_FREEZE_LEVEL, before modifying a position, and keeps the current level rather than sending a rejected request.
• Spread widens exactly where this strategy enters. The extremes it fades often occur in thin conditions. A test run on average spread overstates the edge of an EA whose entries cluster where spread is worst. Test with realistic or real-tick spread, and read the modelling quality before the profit.
• Holds are longer than the reputation implies. A reversion that takes days pays SYMBOL_SWAP_LONG or SYMBOL_SWAP_SHORT every night, and triple-swap day differs by broker.

Failure modes: how a mean-reversion EA loses money

• The stretch keeps stretching. The normal way this strategy loses. Every entry is against a move that does not return, and because the rule keeps re-qualifying, the losses arrive in a block.
• The mean moves. A repriced pair has a new average, and the EA keeps measuring distance from the old one.
• The stop is removed to avoid the loss. Widening a stop, or replacing it with an average-down, converts a defined-risk strategy into a grid with a good story. The equity curve improves immediately and the tail risk becomes unbounded.
• Correlated positions are counted as diversification. One reversal rule on two correlated pairs is the same bet twice. The Dual-Pair basket’s -36.38% drawdown is what that looks like at stock settings.
• Exits are tuned to rescue an entry. When no stop-and-target combination wins in-sample, searching harder only finds a fitted result.
• Cost is measured on the wrong bar. Entries cluster at extremes where spread is widest, and multi-day holds pay swap. Both are invisible in a test configured with fixed average spread and no financing.

How to build and stress-test a mean-reversion EA in AIStrategyMiner

You do not need to write MQL5 to assemble the machine described above. A deviation measure, an oscillator confirmation, a regime gate, a target at the mean and a stop beyond the extreme are all blocks in the no-code Builder, and the reversal templates above arrive pre-wired. The fastest useful test is to load RSI Reversal and MA Pullback Buy side by side. The most visible difference between them is the trend filter, though the threshold, direction and exits differ too — so add an MA Distance Filter to RSI Reversal and change nothing else.

The .ex5 compiled from your own flow is free. The .mq5 source is part of Pro, a one-off purchase with no subscription — see pricing if you want to edit the generated code. The site’s methodology sets out how backtest results should be read.

Mean reversion versus trend

	Mean reversion	Trend-following
What it assumes	A stretched price returns to its average	An established direction continues
Entry relative to the move	Against it, at an extreme	With it, after confirmation
Loses when	The market trends and the stretch keeps stretching	The market ranges and moves fold back
Typical exit	Fixed target at or near the mean	Trailing stop or signal reversal
Worst case	A run of positions held against a move that does not return	A long series of ordinary losses
Complexity	Intermediate — the two exit distances carry the risk	Beginner — few inputs, hard to sit through

Neither is safer in general: one needs stillness, the other needs movement. That is an argument for running both, with one caveat the table cannot show. A mean-reversion EA and a trend EA with a fixed take-profit both depend on price stopping where the rule says it should, and a reversal entry behind a trend filter is already half of each. Check what each exit does before assuming two EAs have diversified anything.