Running Freqtrade Backtests and Understanding Their Limits
Summary
This documentation page explains how to run Freqtrade backtests on historical OHLCV data, select a strategy, timeframe, date range, trading pairs, starting balance, stake settings, fees, and output format, and compare multiple strategies in one run. It describes exporting trades or signals, reviewing breakdowns, and using finer detail candles to model trade timing within a larger candle. Dynamic stake sizing can split the wallet across the maximum number of open trades, allowing early profits to affect later stakes.
The page stresses that suitable historical data must be available and that fees are included in profit calculations. Dynamic pairlists can reflect current rather than historical market conditions, weakening reproducibility; a pre-generated static pairlist is suggested for consistent comparisons. Backtest output can help evaluate a strategy, but it does not establish that it will succeed live. The examples are command usage and sample output, not evidence that any named strategy is profitable.
Key ideas
- Backtests require historical candle data for the chosen exchange, pairs, and timeframe.
- The command options control the test period, fees, stake, starting balance, trade limits, and exported results.
- Multiple strategies can be compared in one run when they use the same timeframe.
- Dynamic pairlists may use current market conditions and make historical results less reproducible.
- Finer detail candles can change simulated trade timing and how quickly trade slots become available.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.