Understand the Numbers Before You Compare Them
A clear explanation of how performance is measured, what each data category means, and the limitations you should know before evaluating any strategy.
Every Result Is Clearly Labelled
Trader AI categorises all performance data so you always know what type of result you are viewing.
Live Performance
Trading activity executed with real funds in live market conditions. Reflects actual spreads, commissions, slippage and market impact. Subject to all risks of real trading.
Paper-Traded Performance
Generated with simulated funds in a simulated environment. May not fully capture slippage, liquidity constraints or psychological factors. Results may differ materially from live trading.
Backtested Performance
Generated by applying trading rules to historical market data. Inherently hypothetical. May incorporate hindsight bias. Does not reflect live execution conditions or changing market regimes.
Illustrative Data
Sample data created solely to demonstrate platform features. Not real performance of any kind. Used in demos, examples and onboarding materials. Should not be interpreted as actual results.
How Performance Metrics Are Calculated
Net Return
Net return is calculated as the percentage change in account equity over a defined period, after deducting trading costs including spreads, commissions and fees. For backtested and paper-traded results, costs are estimated based on configured parameters. For live results, costs reflect actual execution. Returns are calculated on a time-weighted basis where multiple deposits or withdrawals occur.
Maximum Drawdown
Maximum drawdown measures the largest peak-to-trough decline in account equity over the measurement period. It is expressed as a percentage of the peak equity value before the decline began. Drawdown is calculated on a continuous basis using closing equity values, which may not capture intra-period extremes. A large maximum drawdown relative to net return may indicate higher risk per unit of return.
Win Rate
Win rate is the percentage of completed trades that closed with a profit, excluding breakeven trades. It is calculated as: (Winning Trades ÷ Total Completed Trades) × 100. Win rate alone does not indicate whether a strategy is profitable — a strategy with a low win rate can be profitable if average gains significantly exceed average losses. Always evaluate win rate alongside profit factor and average trade metrics.
Trading Costs
All performance calculations include estimated trading costs: spreads (the difference between bid and ask prices), commissions (per-trade fees charged by brokers), slippage (the difference between expected and actual execution price), and overnight financing or swap fees where applicable. Cost assumptions are visible in backtest configurations. Live results reflect actual costs incurred.
Profit Factor
Profit factor is the ratio of gross profit to gross loss across all trades. A profit factor above 1.0 indicates that total profits exceeded total losses. It is calculated as: Total Gross Profit ÷ Total Gross Loss. Higher values indicate more profit generated per unit of loss. This metric helps assess strategy robustness independent of trade frequency.
Strategy Changes and Data Corrections
Strategy parameters may be updated over time. When a strategy is modified, its historical track record prior to the change is preserved with a clear annotation. Data corrections are applied when errors are identified, with the correction date and nature of the change documented. Performance data is recalculated when underlying data is corrected.
Update Frequency and Data Limitations
Update Frequency
Performance data is refreshed based on the data category. Live data updates in real-time or near real-time where supported by data providers. Backtested and paper-traded metrics update when a new simulation is run. Illustrative data is static unless updated for demonstration purposes. Update timestamps are displayed wherever data is shown.
Missing Data
Periods where market data is unavailable are excluded from performance calculations and clearly noted. Strategies with insufficient data for statistical significance are flagged. Metrics based on fewer than 30 trades should be interpreted with caution as sample sizes may not be statistically meaningful.
Performance Methodology FAQs
Backtested results are hypothetical — they show how a strategy would have performed in the past, which may not predict future performance. They are labelled separately so you never confuse historical simulation with live trading results.
Update frequency depends on data provider availability and your plan. Live data is updated in real-time or near real-time where possible. Each data display includes a timestamp showing when it was last updated. Check the data status indicator on any chart or metric.
Not necessarily. A strategy with an 80% win rate can still lose money if the 20% of losing trades are very large. Always evaluate win rate alongside profit factor, average gain versus average loss, and maximum drawdown. No single metric tells the full story.
Yes. All performance calculations include estimated spreads, commissions and slippage. For live results, actual costs are reflected. For backtested and paper-traded results, costs are based on configured parameters. Cost assumptions are always visible in the backtest configuration.
See Performance Methodology in Action
Apply these standards to real strategy comparisons on the AI Trader Leaderboard.