Learn how maximum favorable excursion and maximum adverse excursion can reveal stop, entry, and exit problems—and turn those observations into testable rules.
Quick Answer
Maximum favorable excursion measures the furthest a trade moved in your favor while it was open. Maximum adverse excursion measures the furthest it moved against you. Together, MFE and MAE can reveal whether stops are routinely too tight, entries create unnecessary heat, or trades surrender substantial open gains before exit.
Use these measurements across a sufficiently large, comparable trade sample. They diagnose trade paths; they do not identify an ideal exit by themselves. Any proposed change must be retested chronologically with realistic costs and unseen data.
Key Takeaways
- MFE records the best unrealized movement during a trade; MAE records the worst.
- Normalize excursions in R, percentages, or volatility units when comparing different trades.
- Study distributions and percentiles rather than relying only on averages.
- Never set stops from winning trades alone; doing so ignores how the change affects losses.
- Treat excursion patterns as hypotheses, not proof that another exit would have worked better.
- Recalculate the complete strategy after changing stops, targets, sizing, or trailing rules.
Understanding Maximum Favorable Excursion and MAE
For a long trade, price-based excursions are:
MFE = highest price during the trade − entry price
MAE = entry price − lowest price during the trade
For a short trade, the directions reverse:
MFE = entry price − lowest price during the trade
MAE = highest price during the trade − entry price
Both are usually expressed as positive distances. They can be reported in price points, percentages, currency, volatility units, or multiples of initial risk.
Suppose a long position is entered at $100 with an initial stop at $98. Initial risk is $2 per share, or 1R. While the position is open, price reaches $104, falls as low as $98.50, and eventually exits at $101.20.
The measurements are:
- MFE: $4, or 2R
- MAE: $1.50, or 0.75R
- Realized result: $1.20, or 0.6R
- MFE capture:
0.6R ÷ 2R = 30%
The trade captured 30% of its best unrealized gain. That warrants investigation, but one observation does not prove the exit was defective. A rule intended to retain more of this move could cut other winners short or increase turnover and costs.
MFE is not available profit
MFE identifies the best price observed while the position was open. That point becomes obvious only after the path unfolds. A trader may not have had a valid signal, sufficient liquidity, or executable order at that exact price.
MAE is not automatically the correct stop distance
If successful trades often experience 0.8R of adverse movement, placing the stop beyond 0.8R may appear logical. But a wider stop changes loss size, position sizing, payoff distribution, and which trades remain open. It must be modeled as a new strategy rule.
Publishing a strategy to the Kvants marketplace.
Define Excursions Consistently
Excursion statistics become misleading when measurement policies vary between trades.
Select the price source
Bar highs and lows capture more movement than closing prices, but they do not reveal the sequence of events inside a bar. More granular data can improve sequencing without eliminating every execution uncertainty.
Apply one convention consistently. Do not switch between extremes and closing prices based on which produces a more attractive result.
Define the holding interval
State whether measurement begins at the signal, order submission, or actual fill. End it at the first exit or final exit. Positions with scaling may require a position-level unrealized P&L series rather than a simple entry-to-extreme calculation.
Choose a normalization method
Raw dollar excursions are difficult to compare when prices or position sizes differ. Common methods include:
- R-multiples: Excursion divided by planned initial risk.
- Percentage: Excursion divided by entry price.
- Volatility units: Excursion divided by a predefined measure such as ATR.
- Account percentage: Unrealized change relative to account equity.
The denominator must be fixed using information available at entry. Recalculating it after seeing the result introduces hindsight.
A Step-by-Step Maximum Favorable Excursion Workflow
1. Build a comparable trade sample
Start with trades produced by the same setup and rule version. Record direction, actual fills, initial stop, size, highest and lowest prices while open, exit, costs, and predefined context tags.
Do not combine unrelated strategies. Mean-reversion and momentum trades can have fundamentally different excursion profiles.
2. Calculate normalized MFE and MAE
For each trade, calculate MFE, MAE, and realized R. For straightforward winning trades, you can also calculate:
MFE capture = realized favorable result ÷ MFE
Capture ratios become harder to interpret with partial exits, scaling, gaps, or trades that never moved favorably. A position-level unrealized P&L series is more suitable in those cases.
3. Examine distributions
Review the median, percentiles, range, and shape of the excursion data. Ask:
- How much MAE do typical winners experience?
- Do losses move against the position immediately?
- How often does a trade reach 1R before closing below breakeven?
- Are high-MAE entries concentrated in a particular volatility regime?
- Does low MFE suggest weak entry selection or normal setup behavior?
Outcome groups can assist diagnosis, but live rules must use information available at the decision point. “Apply a wider stop to eventual winners” is not an executable rule.
4. Write an explicit hypothesis
Turn the observation into a rule that can be tested across every eligible trade. For example:
- Move the stop to breakeven only after price reaches 1.5R.
- Replace a fixed target with a predefined volatility-based trailing exit.
- Reject entries when the required stop exceeds the risk budget.
- Exit after a fixed number of bars if MFE has not reached 0.5R.
Specify the threshold, timing, order type, position-sizing response, and treatment of gaps before running the test.
5. Recalculate the complete trade path
Apply the revised rule to all eligible trades in chronological order. Recompute exits, costs, exposure, drawdowns, and position sizes. If a wider stop preserves fixed account risk, position size generally needs to decrease.
Pay particular attention to bars containing both the stop and target. Without sufficient intrabar information, assuming the favorable event occurred first creates optimistic bias.
6. Validate beyond the development sample
Compare the original and revised rules on data not used to design the change. Walk-forward testing can help determine whether the relationship persists through time.
Review expectancy, drawdown, loss size, trade frequency, cost sensitivity, and nearby parameter values. A broad area of similar outcomes is generally more credible than one exceptional threshold surrounded by weak results.
A backtest's equity curve and trade-by-trade log.
Worked Exit-Diagnosis Example
Assume a hypothetical breakout strategy often reaches 1R of MFE but closes many trades near breakeven. This pattern suggests an exit question; it does not prove that 1R should become the target.
Define several candidates in advance:
- Keep the original exit as the baseline.
- Exit the full position at 1R.
- Move the stop to breakeven after 1R.
- Take a partial exit at 1R and trail the remainder with a fixed rule.
Test every candidate on the same chronological data with commissions, spread, slippage assumptions, and consistent order priority. A 1R target might retain more open gains while removing occasional large winners. A breakeven stop might reduce some losses but create more premature exits.
Judge the alternatives by their complete cost-adjusted distributions and validation results—not by selected trades where a revised rule would have helped.
Common MFE and MAE Failure Modes
Designing stops from winners only
Winner MAE shows how much adverse movement successful trades survived. It says nothing about the losses made larger by a wider stop. Test the proposed distance across the complete sample.
Treating MFE as an achievable exit
Each trade’s maximum is visible only in hindsight. Consistently exiting at the historical high or low is not a tradable rule.
Ignoring intrabar ambiguity
A candle can contain both a stop and a target. If their sequence is unknown, favorable assumptions can materially overstate a strategy’s result.
Mixing incompatible conditions
Excursions can change with strategy version, volatility, liquidity, and session. Segment only when there is a defensible reason, and avoid searching many combinations until an appealing result appears.
Ignoring sizing and costs
Tighter exits can increase turnover. Wider stops can require smaller positions under a fixed-risk model. Both effects can reverse an apparent improvement.
Testing Excursion-Based Rules With Kvants
After excursion analysis produces a precise hypothesis, Kvants Studio can translate the idea into editable, auditable strategy logic. Candidate stops, targets, time exits, and trailing conditions can then be evaluated as explicit rules rather than applied selectively in hindsight.
Kvants backtests run on NautilusTrader’s event-driven engine. Parameter sweeps can compare nearby thresholds, while walk-forward and crisis-stress validation can examine behavior beyond the period that inspired the rule. Pine Script v6 export and controlled paper or live workflows support later implementation stages.
The objective is not to select the highest historical result. It is to determine whether the rule remains coherent after realistic sequencing, costs, parameter variation, and unseen data. The Kvants research guides cover related backtesting and validation concepts.
A strategy laid out end to end in the Kvants editor.
Frequently Asked Questions
What is a good maximum favorable excursion?
There is no universal benchmark. MFE depends on the strategy, market, volatility, holding period, and measurement unit. Compare it with initial risk, realized results, and the strategy’s intended payoff structure.
Should a stop be placed beyond the average MAE of winning trades?
Not automatically. That method excludes losing trades and can create a wider, more expensive stop. Test the distance across all eligible trades while recalculating position size and account risk.
What does high MFE but a small realized gain mean?
It may indicate that the exit gives back substantial open gains. It can also be an intentional feature of a trend-following strategy designed to retain exposure for rare large moves. Examine expectancy and drawdown before changing the rule.
Should MFE and MAE use candle highs and lows?
They can, provided the convention is consistent. Candle extremes do not reveal intrabar order, so ambiguous stop-and-target events require conservative assumptions or more granular data.
How many trades are needed for excursion analysis?
No fixed count guarantees a reliable conclusion. The sample should contain enough comparable trades and market conditions to estimate a distribution while preserving separate data for validation. Small samples are better used to generate questions than optimize thresholds.
Risk Note
This article is educational and is not investment advice. MFE and MAE describe historical trade paths; they do not predict future behavior. Backtested performance does not guarantee future results. Liquidity, slippage, costs, data quality, and changing market conditions can cause live outcomes to differ materially from research results. Kvants is a research tool, not an investment adviser, and does not guarantee performance.
