How to read Freqtrade backtest results: what the numbers actually mean
Plain-English tour of Freqtrade backtest metrics, what traders misread, and which numbers matter when you graduate to robustness review.
Freqtrade backtest results interpretation is one of the highest-intent searches in the Freqtrade ecosystem. New traders see a wall of fields; experienced traders know only a few numbers decide whether the run deserves a second validation pass.
This page is a practical tour: what to read first, what traders misread, and how to graduate from "pretty equity curve" to freqtrade strategy validation.
Read profit and risk together first
Start with net profit (or percent) versus drawdown. A strategy that doubles account in a backtest but does so with a 90% underwater path is a different product than the same net profit with a 15% drawdown.
Ask: "Is the payoff worth the risk path on my account size and psychology?"
Trade count: the gatekeeper metric
Freqtrade edge ratio discussions often skip the obvious: if you have 12 trades in a year, your Sharpe and profit factor are mostly noise (How many trades).
Profit factor and expectancy (sanity, not proof)
Freqtrade profit factor is a useful filter. It is not proof of future performance.
Pair expectancy per trade with average win / average loss so you do not get fooled by win rate alone (Win rate vs profit factor).
Sharpe in Freqtrade
Freqtrade sharpe ratio prints are easy to quote and easy to misread. Sharpe is unstable on short samples and blind to tail risk (Sharpe limitations).
Win rate without structure
High win rate with tiny wins and rare huge losses can look "safe" until it is not. Always read win rate distribution, not only the headline percent.
What traders misread most often
- Backtest too optimistic on fees and slippage (Cost drag)
- Cherry-picked date ranges after peeking at results (Data snooping)
- Hyperopt winners taken as final truth (Hyperopt)
Next step: beyond the summary table
When you are past "first backtest," export to Kiploks for walk-forward and robustness framing (Beyond backtesting, Integration).
FAQ
What is a good profit factor? Context-dependent. Compare against costs, sample length, and whether results survive time splits.
Why does my Freqtrade backtest look better than live? See Why your Freqtrade backtest is great but live trading fails.