MFE and MAE in a trading journal: measuring favorable and adverse excursion without hindsight bias

MFE and MAE in a trading journal: measuring favorable and adverse excursion without hindsight bias

Final profit hides the path

A trade's net result tells you where it finished, not the complete route taken while it was open. Maximum Favorable Excursion, or MFE, describes the largest observed favorable movement from entry during the holding period. Maximum Adverse Excursion, or MAE, captures the largest observed movement against the position. These are retrospective measurements tied to a defined market-data series. A profitable trade might first have endured a deep adverse swing; a losing trade might temporarily have been far ahead. Recording both can improve questions about execution, stop placement and exit discipline. It cannot demonstrate that a trader could have known an eventual extreme in advance or that the same path will recur.

Define the trade before measuring it

The calculation needs a direction, actual entry price, exact opening timestamp and closing timestamp. For a long position, prices above entry can contribute to favorable excursion and lows below it to adverse excursion; for a short position the definitions reverse. A single OHLC candle containing the entry may include a high or low reached before the order existed. Assigning that entire candle's extreme to the trade can manufacture an opportunity that was never available. When intrabar information is missing, disclose the approximation or mark the observation unknown. Clock alignment, exchange sessions and the distinction between prices and fills are part of statistical accuracy, not optional presentation details.

Use the executable side of the market

A chart based on bid prices does not necessarily reveal the ask price paid on entry into a long position. Excursions estimated using mid-price or last-trade highs can differ from what was executable with a particular broker. Variable spread, commissions, financing and slippage do not vanish because a chart touched an attractive high. Clearly identify whether MFE and MAE use observed high/low prices or bid/ask histories representing actual closing conditions. Do not display theoretical maximum favorable price movement as money that could definitely have been captured. Data frequency changes accuracy too: hourly bars cannot settle the sequence of price extremes inside one hour.

A transparent illustrative long trade

Imagine a hypothetical long entered at a price of one hundred, with the highest observed print at one hundred and four and the lowest at ninety-eight while the position was open. The favorable excursion is four price units and the adverse excursion is two, before costs and executable quote differences. If the position eventually closes at one hundred and one, the final gain of one unit omits those extremes. This does not prove that a close at one hundred and four would have been easy or possible with the relevant order type and liquidity. The data show a historical maximum, not the trader's ability to anticipate it. For comparative review, express the measures in initial-risk multiples or account currency when their calculation is valid.

Keep the original risk denominator

Dividing excursions by the original distance to stop can produce comparable R multiples, provided risk was recorded consistently at entry. If the stop is moved afterwards, keep the starting risk and track each amendment separately; do not rewrite the denominator when the result becomes known. Otherwise the apparent efficiency of a trade may improve merely because its original plan was edited. Currency conversion, position volume and commissions also matter when assessing results across symbols. R normalization does not automatically account for execution costs or different broker contracts. Specify whether the measure refers to theoretical chart distance or actual money exposure.

Partial closes and changing volume

An MT5 position can generate several execution deals and partial closes. If half the position closed before a later high was reached, it would be misleading to attribute that eventual favorable movement to the entire original volume. You can define excursion for each surviving volume segment or for the overall trading idea, but the method must be applied consistently. Netting accounts can alter one consolidated position while hedging accounts may retain several concurrent tickets. A blind count of deals will not reproduce economic exposure. Preserve position identifiers, volume changes, timestamps and direction before presenting a dashboard. A single neat statistic is not worth losing the actual sequence of capital at risk.

Hindsight bias is the central trap

A large observed MFE does not demonstrate that a wider profit target was reasonable before entry. The distribution depends on which trades were selected, prevailing conditions and when positions were closed. Studying only winners creates selection bias. Choosing the historically perfect exit after observing each price path produces an impressive but untradeable backtest. A legitimate improvement defines a proposed rule using one historical sample, then tests it on later data not used for fitting. The review must distinguish evidence, new hypothesis and out-of-sample performance. No result guarantees future profit; the objective is to reduce unsupported decisions and identify which assumptions survived independent observations.

MAE and stop placement require care

If many winning trades first experience substantial MAE, that is not an automatic reason to widen every stop. Ask whether entry occurred before confirmation, whether the setup tends to test a liquidity region and how much losing trades would cost under a wider limit. Moving a stop further away without reducing volume increases monetary exposure and may breach portfolio limits. A distribution of historical MAE can raise questions about timing, but should not dictate a new threshold from a handful of favorable examples. Stops should remain linked to a defensible technical invalidation and actual brokerage conditions. Stress cases with gaps or unusual liquidity need separate treatment.

MFE and exit management

If observed MFE regularly exceeds final captured profit, investigate whether a trailing rule, partial exit or time-based exit might improve the process. Do not assume the answer is yes. A rule that protects some gains may cut short large trends and incur more transaction costs. Evaluate complete samples, not only visually dramatic cases. Simulations should use data available at each decision point and avoid selecting a later-known peak as the trigger. A fixed take-profit also has to be tested against trades that never reach the threshold or hit their stop first. A retrospective metric is not an automatic instruction to close.

Segment by session and setup

Compare excursions across documented setups, entry sessions, timeframes and volatility regimes. Breakouts around high-impact news can have very different paths from calm-market pullbacks. Do not claim a particular session causes better MFE without sufficient synchronized observations. Missing candles, spread histories or contradictory timestamps should reduce the confidence of the result. A few selected trades cannot support precise percentiles with meaningful stability. Show sample sizes alongside summaries so a user can tell the difference between an emerging pattern and an artifact of sparse data. A robust system is capable of saying that a question cannot yet be answered.

Behavioral review in Disciply

Disciply can connect journal information and behavioral tags with price excursions, but a psychological state cannot be inferred reliably from a high/low pair. An impulsive entry and a disciplined entry may happen to have identical MFE. Record intended plan, original stop, confirmations, changes, session and reason for exit explicitly before examining performance. Then investigate recurring decisions using adequately sized groups. A winning outcome after violating a rule remains a process issue, while a losing planned trade is not proof of indiscipline. The goal is to improve observable behavior rather than grade a trader solely by a monetary result.

Implementation checklist and references

Preserve account identifier, broker server, position ID, order direction, entry and exit times, evolving volume, original stop, appropriate OHLC history and executed costs. Validate candle continuity and state when precise excursion is unavailable. MetaTrader explains deals at https://www.metatrader5.com/en/terminal/help/trading/performing_deals and history at https://www.metatrader5.com/en/terminal/help/trading_advanced/history_report . An associated Disciply guide is https://disciply.it/blog/diario-mt5-riconciliazione-deal-posizioni-chiusure-parziali-2026 . The method is educational, not investment advice or a promise that optimized exits will generate profit.