What is backtesting?

Backtesting simulates a trading strategy by applying predefined rules to historical market data. Data quality, fee assumptions and the order-fill model affect the reported result. The exercise describes a past scenario rather than predicting future returns.

Example: gross profit versus modeled costs

Suppose a hypothetical strategy calculates a 60 USDT gain before fees. With 20,000 USDT of total executed turnover and an assumed 0.1% fee, commission is 20 USDT, leaving 40 USDT. That figure excludes slippage or funding if those costs were not modeled. Trading turnover is not the same measure as starting capital.

What does a backtest fail to establish?

Selecting parameters only because they look best in one historical period can overfit that data. Evaluate a separate period that was not used to choose the settings. Using a later price that was unavailable at the decision time creates look-ahead bias and can artificially improve results.

Paper trading follows newly arriving prices with simulated orders; backtesting uses historical data. Neither perfectly reproduces live queue position, liquidity or outages. Our trading bot platform comparison also examines which components are excluded from particular testing tools.