Trading performance metrics should tell you four different things: whether the sample made money, how that result was produced, how much path risk it required, and whether the process was repeatable. No single number can answer all four questions.

A useful core set is: net result, win rate, average win and loss, expectancy, profit factor, R-multiple distribution, maximum drawdown, and sample size. Add Sharpe or Sortino only when you have a consistent return series and understand the assumptions. Add setup, session, MAE/MFE, duration, and rule-compliance breakdowns when your underlying trade log supports them.

The important rule is not to chase a universal target such as “60% win rate,” “2:1 R:R,” “profit factor above 1.5,” or “drawdown below 10%.” The same value can mean different things for different strategies, samples, instruments, costs, and risk levels.

Key Takeaways

Win rate is incomplete. It ignores the size of wins and losses.
Expectancy and profit factor summarize different views of the same trade outcomes, but neither shows loss sequence or drawdown by itself.
Maximum drawdown is historical path risk, not a guaranteed worst-case future loss.
Planned risk-reward and realized payoff are different metrics. Do not mix them.
R-multiples require a frozen initial-risk definition. If initial risk was not recorded before the trade, the metric can be misleading.
Sharpe ratio is not a universal trader score. It depends on the return series, benchmark or risk-free assumption, volatility estimate, sampling interval, and return distribution.
Sample size belongs on the dashboard. A strong metric from a small or cherry-picked sample is weak evidence.
Costs and execution belong inside the numbers. Gross chart results are not the same as net trading performance.
Segment analysis is useful only when tags are defined consistently. Repeatedly slicing data until one subgroup looks good creates hindsight bias.

Last reviewed: August 21, 2026.

What This Page Owns

This guide owns the calculation and interpretation of aggregate trading-performance statistics.

Related pages answer different questions:

Trading Journal Guide — what raw fields to record and how to maintain the journal.
Trading Journal Review System — how to organize weekly, monthly, and periodic review.
Post-Trade Review — how to analyze one completed trade without turning one outcome into a strategy conclusion.
How to Backtest a Trading Strategy — how to construct and test a historical strategy sample.
Risk-Reward Ratio Explained — planned reward relative to planned risk before entry.
Risk Management and Position Sizing — how risk budgets, sizing, leverage, execution, and portfolio exposure are controlled.

A journal supplies the data. A backtest supplies one type of historical sample. This page explains how to read the resulting numbers without treating any one statistic as proof of a durable edge.

Start With Data Quality Before Calculating Metrics

A performance dashboard is only as reliable as its trade records.

Before calculating statistics, define what one row represents and record enough information to reconstruct the result. A useful trade record can include:

strategy or setup version;
instrument;
direction;
entry and exit timestamps;
entry and exit fills;
quantity;
commissions and fees;
estimated or recorded slippage;
financing or borrow costs when relevant;
initial stop or other initial-risk definition;
realized P&L;
setup, session, and market-context tags;
whether the trade followed the written rules.

Do not mix gross and net results

A strategy can look attractive before costs and weak after them. Decide whether every metric is calculated from:

gross trading result; or
net result after the costs you can reasonably measure.

For real-account review, net results are usually more useful. For research, it can be useful to show both so you can see how much of the gross edge is consumed by friction.

Separate deposits and withdrawals from trading performance

If outside cash is added to or removed from an account, the simple change in account balance is not automatically the trading return.

For example, an account that rises from $10,000 to $15,000 after a $5,000 deposit did not generate a 50% trading return. Keep external cash flows separate from strategy P&L, and use a consistent return methodology when comparing periods.

A Practical Performance-Metric Stack

Instead of asking for “the best metric,” organize metrics by job.

Layer	Metrics	Main question
Outcome	Net P&L, return	What happened over this sample?
Frequency and payoff	Win rate, average win, average loss, payoff ratio	How were wins and losses distributed?
Edge summary	Expectancy, profit factor, average R	Did the historical sample produce more than it lost under these rules?
Path risk	Maximum drawdown, losing streaks, recovery time	How difficult was the path?
Risk-adjusted return	Sharpe, Sortino or another declared measure	How did return compare with the chosen risk measure?
Diagnostics	Setup/session tags, MAE/MFE, duration, compliance	Where did the result come from?
Evidence quality	Trade count, date range, market regimes, cost assumptions	How much confidence should you place in the statistics?

The layers matter because two strategies can produce the same net profit with very different drawdowns, trade counts, concentration, and execution burdens.

Metric 1: Net P&L and Return

Net P&L is the closed trading result after whatever costs your calculation includes.

net P&L = gross trading P&L - commissions - fees - financing - other included costs

Return is usually expressed relative to a defined capital base:

period return = (ending value - beginning value - external net cash flows) / adjusted capital base

The exact return methodology can become more complex when cash flows occur during the period. The important point for a personal trading dashboard is to avoid treating deposits as trading gains or withdrawals as trading losses.

What net P&L does not tell you

Net P&L alone does not show:

how much capital was at risk;
whether one outlier produced most of the gain;
how deep the equity curve fell along the way;
whether the result came from a stable setup or rule-breaking;
whether a comparable passive benchmark would have produced a different result;
whether the same process worked outside the selected period.

That is why P&L is an outcome metric, not a complete performance diagnosis.

Metric 2: Win Rate

Win rate is the proportion of resolved trades classified as winners.

win rate = winning trades / included trades

Define how breakeven trades are handled before calculating the rate. Possible conventions include excluding exact breakevens from the denominator or treating them as a separate third outcome. Either can work if you use the same rule every time.

Why win rate cannot be read alone

A high win rate can lose money if losses are much larger than wins. A lower win rate can make money if the payoff distribution compensates for more frequent losses.

Therefore pair win rate with at least:

average win;
average loss;
expectancy; and
profit factor.

There is no universal “good win rate” for scalping, day trading, swing trading, or trend following. The rate depends on the strategy's exit logic, payoff shape, costs, and sample.

Metric 3: Average Win, Average Loss, and Realized Payoff Ratio

Calculate average winning and losing outcomes from realized results, not planned targets.

average win = gross or net profit from winners / number of winners
average loss = absolute gross or net loss from losers / number of losers

realized payoff ratio = average win / average loss

This ratio is different from the planned risk-reward ratio written before entry.

A plan may target 2R but realize 0.8R winners because trades are cut early, gaps occur, partial exits change the outcome, or targets are rarely reached. Performance analysis should preserve both fields when possible:

planned R:R — what the trade intended before entry;
realized payoff — what the sample actually produced.

Neither has a universal target. The correct pairing depends on win frequency and costs.

Metric 4: Expectancy

Expectancy estimates the average historical result per trade in the sample.

When average loss is stored as a positive magnitude:

expectancy
= (win probability × average win)
- (loss probability × average loss)

Suppose a sample has:

win rate: 40%;
average net win: $300;
loss rate: 60%;
average net loss: $150.

Then:

expectancy = (0.40 × 300) - (0.60 × 150)
           = 120 - 90
           = $30 per trade

That means the observed sample average is $30 per included trade under those definitions. It does not mean the next trade is expected to make exactly $30 or that the true future expectancy is known.

Expectancy is an estimate, not a promise

Historical expectancy can move substantially when:

the sample is small;
one or two outliers dominate results;
market regime changes;
costs change;
strategy rules drift;
execution quality changes;
the sample was selected after seeing the outcome.

Track the number of observations and strategy version beside expectancy. A number without its sample context is easy to overinterpret.

Metric 5: Profit Factor

Profit factor compares total gross profit from winning trades with the absolute total gross loss from losing trades.

profit factor = gross profit / absolute gross loss

If a sample has $8,000 of winning-trade profits and $6,000 of losing-trade losses:

profit factor = 8,000 / 6,000 = 1.33

A value above 1 means gross profits exceeded gross losses in that sample under the chosen cost convention. A value below 1 means the reverse.

Do not turn that identity into an unsupported universal threshold such as “1.5 is good” or “2.0 is professional.” A profit factor of 1.3 from a large, stable, cost-adjusted out-of-sample record can be more informative than 3.0 from a tiny hand-picked sample.

Profit factor does not tell you:

when the losses occurred;
how deep drawdown became;
whether one large winner dominated gross profit;
how much capital was required;
how many observations produced the ratio.

Use it with drawdown, sample size, and distribution information.

Metric 6: R-Multiples and Average R

An R-multiple normalizes a trade by the amount of initial planned risk.

realized R = realized net trade P&L / initial planned risk amount

If $100 was the pre-entry planned loss unit and the trade closes at +$180 net:

realized R = +1.8R

If it closes at -$75:

realized R = -0.75R

R-normalization can make trades of different dollar sizes easier to compare, but only if the denominator is defined consistently.

Freeze the denominator before entry

Do not recalculate “initial risk” after the trade to make the R result look better. Record:

initial entry assumption;
initial invalidation/stop reference;
position quantity;
initial planned dollar risk.

If a trade has no defensible initial-risk record, it may be better to omit the R-multiple than invent one afterward.

Metric 7: Maximum Drawdown

Maximum drawdown measures the largest observed peak-to-trough decline in an equity curve over the sample.

For each historical peak:

drawdown = (current equity - prior peak equity) / prior peak equity

Maximum drawdown is the largest decline in magnitude observed across the period.

For example, if an equity curve reaches $20,000, falls to $16,000 before making a new high, and no deeper peak-to-trough decline occurs:

maximum drawdown = (16,000 - 20,000) / 20,000 = -20%

Historical maximum drawdown is not a risk ceiling

A strategy that historically drew down 8% is not guaranteed to stay within 8% in the future. Future sequences can be worse, liquidity can change, slippage can increase, and the strategy can stop working.

Drawdown is therefore best used as:

a description of observed path risk;
a comparison point across strategy versions under the same assumptions; and
an input to risk-capacity decisions.

It is not a guarantee of the maximum future loss.

The asymmetric recovery arithmetic is still useful:

Drawdown	Gain required to recover to prior peak
10%	11.1%
20%	25.0%
30%	42.9%
50%	100.0%

This is arithmetic, not a forecast of how long recovery will take.

Metric 8: Losing Streaks and Recovery Time

A maximum losing streak counts the largest number of consecutive losing trades in the sample. Recovery time measures how long an equity curve remains below a previous peak before reaching a new high.

These metrics add sequence information that profit factor and expectancy do not contain.

But avoid reading psychology directly from a streak statistic. A five-trade losing streak may be normal for one strategy and unusual for another. The useful comparison is against the tested distribution for the same strategy version, not against an internet rule.

Metric 9: Sharpe Ratio

The Sharpe ratio is a risk-adjusted return measure based on excess return relative to a benchmark or risk-free return divided by the variability of that differential return.

A common historical form is:

Sharpe ratio = average periodic excess return / standard deviation of periodic excess returns

William F. Sharpe's own discussion distinguishes ex ante and ex post versions and emphasizes that historical values are estimates, not guaranteed forecasts. The calculation also depends on the return interval and the benchmark or risk-free series used.

Do not calculate Sharpe from a random mixture of trade P&Ls

For a personal trading dashboard, decide on a consistent return series first—for example daily or monthly account returns with external cash flows treated appropriately. Then keep the periodicity and annualization convention consistent when comparing strategies.

Why a universal Sharpe target is weak guidance

CFA Institute's discussion of the metric notes that Sharpe compresses performance into mean and standard deviation and can be incomplete for asymmetric or non-normal return distributions.

A strategy can have:

attractive average return;
modest measured volatility;
and still contain negative skew, tail risk, illiquidity, or path risk not summarized well by standard deviation.

So read Sharpe alongside maximum drawdown, the shape of returns, liquidity/execution assumptions, and the actual strategy.

Metric 10: Sortino Ratio—Useful, but Define the Convention

Sortino-style measures replace total volatility with a downside-risk measure. This can be useful when upside variability should not be treated the same way as downside variability.

However, implementations can differ in:

target or minimum acceptable return;
downside-deviation formula;
sampling interval;
annualization.

Do not compare two Sortino ratios unless they were calculated using compatible conventions and return data.

Metric 11: MAE and MFE

Maximum adverse excursion (MAE) measures how far a trade moved against the position while it was open. Maximum favorable excursion (MFE) measures how far it moved in the favorable direction before exit.

These can answer questions such as:

Do winning trades routinely approach the stop before recovering?
Are losers often profitable first and then allowed to reverse?
Are exits consistently leaving a large amount of favorable excursion unrealized?
Does a proposed tighter stop remove many eventual winners?

But MAE/MFE require intratrade path data. If your dataset contains only entry and exit values, you cannot reconstruct them accurately. Bar-based data can also hide the sequence of events inside a candle.

Do not turn MFE into “maximum profit you should have captured.” That is hindsight. Use it to compare frozen exit-rule versions, not to judge every trade against the best possible exit after the fact.

Metric 12: Duration and Time Exposure

Trade duration can be measured in minutes, bars, sessions, or days.

Useful comparisons include:

duration of winners vs. losers;
time to first favorable move;
time to stop or target;
total market exposure;
strategy results by holding-period bucket.

There is no universal rule that losers must close faster than winners. A mean-reversion strategy, option strategy, trend strategy, and event-driven trade can have very different time distributions.

Use duration to test the strategy's own assumptions—for example, whether a setup loses usefulness after a specified number of bars—not to impose a generic timing rule.

Metric 13: Rule Compliance and Execution Error Rate

Performance statistics describe outcomes. A process metric asks whether the trader actually executed the tested rules.

A simple compliance rate can be defined as:

rule compliance rate = compliant trades / reviewed trades

You can also count specific error types:

early entry;
late chase;
oversizing;
moved stop;
unplanned add;
unplanned exit;
trade taken outside the strategy version;
missed valid setup.

Do not treat a profitable rule violation as evidence that the violation was good. Tag outcome and process separately.

This is the bridge between aggregate performance analysis and the Post-Trade Review.

Segment Performance Without Data Mining Yourself

A blended account metric can hide important differences. It can be useful to calculate expectancy, profit factor, average R, or drawdown by:

strategy/setup;
instrument;
session;
long vs. short;
volatility or market-regime label;
rule-compliant vs. non-compliant trades.

But segmentation creates a new risk: the more slices you inspect, the easier it is to find an impressive subgroup by chance.

Use these controls:

Define tags before reviewing the outcome when possible.
Keep strategy versions separate.
Display trade count next to every subgroup metric.
Do not promote a setup because of five unusually good trades.
Retest a hypothesis on later or out-of-sample data.
Keep cost assumptions consistent across compared groups.

The Backtesting Guide owns the deeper historical-testing workflow, including look-ahead and selection-bias controls.

Sample Size: Put the Denominator on the Dashboard

There is no universal number of trades that makes a strategy “statistically proven.” The amount of evidence required depends on:

variability of outcomes;
size of the estimated edge;
dependence between trades;
number of strategy variants tested;
number of subgroup comparisons;
stability across different conditions.

At minimum, show:

trades: 83
period: 2025-10-01 to 2026-08-01
strategy version: v3
markets: ES, NQ
cost model: commissions + fixed slippage assumption

That context is often more valuable than adding another ratio.

Use rolling metrics to see instability

A single lifetime expectancy can hide change. Compare metrics over predefined rolling windows or strategy versions—but do not react to every short-term fluctuation.

A useful dashboard can show:

full-sample result;
recent fixed-window result;
number of observations in each;
whether the strategy rules changed.

If the recent window deteriorates, investigate. Do not automatically conclude the edge is gone from one noisy window.

A Minimal Trading Performance Dashboard

For many discretionary traders, start with this set:

Metric	Why keep it	Pair it with
Trade count	Shows evidence quantity	Date range and strategy version
Net P&L / return	Shows sample outcome	Costs and cash flows
Win rate	Shows outcome frequency	Avg win/loss
Avg win / avg loss	Shows payoff shape	Win rate
Expectancy	Summarizes average trade outcome	Sample size and distribution
Profit factor	Summarizes gross profit vs. gross loss	Drawdown and concentration
Average/median R	Normalizes by initial risk	Frozen risk definition
Max drawdown	Shows observed path risk	Recovery time and equity curve
Compliance rate	Shows process consistency	Error tags

Add Sharpe, Sortino, MAE/MFE, exposure, turnover, and more detailed distributions when you have the data and a specific question they answer.

Worked Example: Read the Metrics Together

Consider two hypothetical strategy samples. These numbers are illustrations only.

Metric	Strategy A	Strategy B
Trades	40	180
Win rate	70%	42%
Avg win	$80	$260
Avg loss	$220	$140
Expectancy	-$10	$28
Profit factor	0.85	1.34
Maximum drawdown	-7%	-12%

Strategy A has the higher win rate and shallower historical drawdown, but the sample loses money because its losing trades are much larger than its winners.

Strategy B has a lower win rate and deeper historical drawdown, but its sample has positive expectancy and profit factor above 1.

Does that prove Strategy B is a good live strategy? No. You still need to know:

whether the 180 trades were genuinely out-of-sample or cherry-picked;
whether costs and slippage are realistic;
whether a few outliers drive the result;
whether the strategy is stable across relevant conditions;
whether the observed drawdown fits the trader's risk capacity;
whether the rules can be executed consistently.

Metrics help ask better questions. They do not remove uncertainty.

Common Performance-Metric Mistakes

Mistake 1: Using planned R:R as realized performance

A target of 2R is not an average realized win of 2R. Track both separately.

Mistake 2: Calling a threshold “good” without context

A 50% win rate, 1.5 profit factor, 10% drawdown, or 1.0 Sharpe ratio does not have the same meaning across every strategy and sample.

Mistake 3: Ignoring transaction costs

High-turnover strategies are particularly sensitive to spread, fees, slippage, financing, and market impact.

Mistake 4: Treating maximum historical drawdown as the worst possible future loss

It is only the largest drawdown observed in the tested history.

Mistake 5: Annualizing a short or inconsistent return series

Annualized metrics can create false precision when the underlying sample is short, unstable, or measured at inconsistent intervals.

Mistake 6: Comparing Sharpe ratios calculated differently

Risk-free rate, return frequency, annualization, and data treatment need to match.

Mistake 7: Optimizing the dashboard instead of the strategy

If you change exits to increase profit factor, reduce size to improve drawdown, and filter setups to improve win rate all at once, you create a new strategy version. Retest it rather than assuming the old evidence still applies.

Mistake 8: Slicing until something looks profitable

Setup, weekday, session, instrument, and market-condition breakdowns are hypotheses. Validate them on later data.

How to Use Metrics in a Review Process

A repeatable sequence is more useful than staring at a dashboard.

1. Verify the dataset

Check missing trades, duplicates, costs, strategy-version labels, and cash flows.

2. Read outcome and path together

Review net result, expectancy/profit factor, and drawdown instead of only P&L.

3. Check concentration

Ask how much of the result came from the largest few winners or losers.

4. Separate process from outcome

Compare rule-compliant and non-compliant trades without rewarding violations that happened to make money.

5. Form one testable hypothesis

For example:

Breakout trades entered after the defined session window appear to have lower net expectancy after costs than the same setup inside the planned window.

6. Test the hypothesis on a separate sample

Do not rewrite the strategy from one month or one subgroup unless the evidence supports it.

The Trading Journal Review System covers the cadence and decision process around these calculations.

What ChartMini Can and Cannot Measure

ChartMini can support historical chart replay and simplified simulated trade/session review. It can help you practice making decisions without seeing future candles and retain records from the simulated workflow.

It should not be treated as a complete broker-grade performance analytics engine. The core replay does not reproduce every input required for live metrics, including:

live bid-ask spread changes;
real order-book queue position;
broker routing and latency;
partial-fill mechanics;
every commission, financing, borrow, or exchange fee;
guaranteed intrabar sequencing from coarse historical bars.

ChartMini also should not be described as automatically calculating an “optimal” expectancy, profit factor, Sharpe ratio, or strategy recommendation from live trading activity.

For a serious performance review, export or maintain a clean trade record from the environment you actually trade, then calculate metrics using clearly documented conventions.

FAQ

What are the most important trading performance metrics?

Start with trade count, net result, win rate, average win/loss, expectancy, profit factor, R-multiples, and maximum drawdown. Add risk-adjusted or diagnostic metrics only when they answer a specific question and your data supports them.

Is win rate the most important trading metric?

No. Win rate only measures how often trades win. It must be interpreted with the size of wins and losses, costs, expectancy, drawdown, and sample size.

What profit factor is good for trading?

There is no universal target. A profit factor above 1 means gross profit exceeded gross loss in that historical sample under the chosen accounting convention, but confidence depends on sample size, costs, stability, drawdown, and concentration.

What maximum drawdown is acceptable?

There is no universal acceptable percentage. It depends on the strategy, leverage, portfolio, financial risk capacity, and how much worse future drawdown could be than the historical sample. Maximum drawdown is descriptive evidence, not a future-loss guarantee.

What is the difference between expectancy and profit factor?

Expectancy estimates the average historical outcome per trade, while profit factor compares total winning-trade profit with total losing-trade loss. Both summarize profitability but neither describes the full sequence or drawdown path.

Is a Sharpe ratio above 1 good?

Do not use 1.0 as a universal cutoff. Sharpe is sensitive to the return series, benchmark or risk-free rate, volatility estimate, annualization, autocorrelation, and return distribution. It is most useful for comparisons made under compatible assumptions and should be read with other risk measures.

How many trades do I need before performance metrics are reliable?

There is no fixed universal count. More observations generally reduce some sampling noise, but reliability also depends on outcome variability, dependence between trades, strategy changes, multiple testing, and market diversity. Always display the sample size with the metric.

Can ChartMini automatically calculate all of these metrics?

Do not assume so. ChartMini's core replay is designed for historical chart and decision practice with simplified simulated records. It is not a substitute for broker-grade execution data or a full statistical analytics platform.

Sources and Verification Notes

William F. Sharpe — The Sharpe Ratio, Stanford University — original-author discussion of ex ante/ex post Sharpe definitions, differential returns, variability, and interpretation limits.
CFA Institute — The Sharpe Ratio and the Information Ratio — practical discussion of Sharpe-ratio construction and limitations, including mean/variance assumptions and asymmetric return distributions.
Charles Schwab — Elements of a Smart Trade Plan — trade-journal recording and periodic review context.

Formulas in this article are educational definitions. Metric conventions can vary across brokers, analytics platforms, researchers, and strategy-testing software. Before comparing two reports, verify that they use compatible definitions, cost treatment, return frequency, cash-flow treatment, and annualization assumptions.

Trading Performance Metrics: Expectancy, Profit Factor, Drawdown, and More