Definition and what performance statistics actually measure
Performance statistics are summary metrics that describe trading outcomes over a specific time period, usually based on recorded trades, account equity, or returns. Examples include measures of average return, drawdown (a peak-to-trough decline), volatility (how much results fluctuate), and risk-adjusted measures that combine return and variability.
In practice, these statistics are computed from inputs such as timestamps, trade direction and size, quoted or filled prices, and then sometimes converted into returns (percentage changes over time). Because they are derived from observed records and chosen formulas, their meaning depends on the data quality and the assumptions built into the calculation.
How the mechanics create uncertainty
A common failure mode is that performance statistics compress complex processes into a small set of numbers.
-
Data and time-window choices. The selected start and end dates determine which trades enter the calculation. If the window includes unusually calm or unusually volatile periods, the resulting metrics can look more stable or more extreme than they would over other periods.
-
Execution and cost assumptions. Returns are affected by trading costs (spreads, commissions, and other fees) and by execution quality (whether trades fill near the intended price). Many performance summaries do not fully isolate the impact of these factors, so two sets of statistics can describe the same strategy differently.
-
Return series construction. Metrics often rely on how equity or balance changes are recorded, how partial closures are handled, and how missing or corrected data is treated. Small differences in these mechanics can change drawdown size or volatility even if the underlying trading is similar.
-
Selection effects. If results are shown only for periods where a provider traded, or if incomplete histories are excluded, the reported statistics can be a biased sample. This is a measurement limitation, not proof that the underlying approach works or fails.
Evidence and example of where statistics can mislead
Consider two accounts with identical strategy behavior but different measurement environments:
- Account A records returns using a longer period that includes high spread regimes.
- Account B records returns using a shorter period that excludes those regimes.
Even if the strategy’s core mechanics are unchanged, Account B may report lower drawdown and smoother performance simply because the sampled market conditions were different. The limitation is not the arithmetic; it is the mismatch between the conditions used to compute the metrics and the conditions that may occur later.
Another example is drawdown interpretation. A drawdown statistic summarizes a decline but does not explain how long recovery took, whether the decline correlated with specific events, or how sensitive it was to execution changes. Two providers can share the same maximum drawdown value while having very different recovery profiles.
Material limitations and failure modes
Performance statistics have several material limitations:
-
No guarantee of future results. Historical relationships between returns and risk measures do not ensure similar behavior later, especially when volatility regimes and market liquidity shift.
-
Hidden assumptions. Calculations depend on what is counted as a trade, how returns are measured, and whether costs are fully reflected. If assumptions differ across reporting, comparisons can be unreliable.
-
Non-stationary markets. Market dynamics can change over time. A method that performed in one environment may underperform in another, even without any change in execution.
-
Provider and reporting differences. Statistics may reflect how a platform aggregates activity, handles deposits/withdrawals, or corrects or filters activity. These differences can affect performance numbers.
-
Survivorship and reporting bias. If only successful or active histories are displayed, statistics can systematically overstate typical outcomes.
-
Jurisdiction and operational factors. The practical experience of trading can vary with legal and operational settings, which performance statistics alone cannot capture reliably.
Verification and next questions to ask
To verify what performance statistics mean for a specific case, check whether the reported metrics are reproducible from the underlying trade or equity records, and identify the calculation inputs and time windows.
Next questions that help reveal limitations:
- What time period is used, and what market conditions dominate it? - Are costs and execution details included consistently? - How are deposits, withdrawals, and adjustments treated in the return series?