Maximum Adverse Excursion (MAE)
Definition
Maximum Adverse Excursion (MAE) is the worst unrealised loss a trade experienced before closing — how far against you the trade went at its worst point. MAE analysis helps optimise stop-loss placement: if winning trades rarely exceed 1.2R adverse excursion, a stop at 1.5R is wasting potential by sitting too wide.
In-depth: Maximum Adverse Excursion (MAE)
Maximum Adverse Excursion (MAE) was introduced by John Sweeney in 'Maximum Adverse Excursion: Analyzing Price Fluctuations for Trading Management' (1996). The metric measures the worst point reached by an open trade — specifically, the largest unrealised loss before the trade closes (whether by stop, target, or other exit).
Formal definition: MAE for a trade = max(open_loss_during_trade), measured in pips, dollars, or R-multiples depending on context.
Diagnostic use: plotting MAE distributions separately for winning vs losing trades reveals stop-loss optimisation opportunities. - If winning trades rarely exceed (say) 0.8R MAE while the stop is set at 1.5R, the stop is too wide — winners would survive a tighter stop, but losers would lose less. Tightening to 1.0R may improve net edge. - If winning trades regularly approach the stop (MAE close to stop distance), the stop may be cutting winners short. Widening the stop preserves more winners but also costs more on losers; this trade-off is the core of stop-loss optimisation.
The critical insight: MAE distributions tell you where your edge actually operates in adverse-price terms. A strategy whose winning trades have MAE < 0.5R is a strategy where most winners go directly in your favour. A strategy whose winning trades have MAE up to 1.2R is one where trades often go against you before recovering.
Combined with Maximum Favourable Excursion (MFE), MAE characterises the trade dynamics: high MAE + high MFE = volatile trades that move around a lot; low MAE + low MFE = trades that exit cleanly without much intraday adverse movement.
Implementation in MQL5 EAs: track each trade's lowest equity point during the holding period. Store MAE alongside trade outcomes for post-hoc analysis. The MetaTrader strategy tester does not natively output MAE — custom logging or third-party analysis tools are needed.
Limitations: MAE analysis assumes future trades will have similar MAE distributions to historical, which depends on regime stability. MAE optimisation can also overfit to historical noise — adjust stops based on MAE patterns only when sample size is large enough (typically 100+ trades) and patterns are stable across out-of-sample data.