Your First Strategy Backtest
Published Jun 22, 2026 · Updated Oct 16, 2026
Start with one small goal: create an experiment, finish a saved run, and check what the strategy did. Where to start: Strategy → Strategy backtest. The backtest replays historical market data in your browser; creating or running one never starts a trading strategy.
Sign in to AlphaFox first.
Run your first backtest
Create a backtest experiment
Open Strategy → Strategy backtest; it opens on the “Backtest experiments” list. Select New backtest. In the “New backtest experiment” dialog, pick a template under “Strategy”. The name is optional and defaults to the strategy name. Select Create, and the experiment’s workspace opens.
One experiment can hold several saved run versions for comparison.
Set this run’s inputs
In the “Parameter explorer”, set:
| Field | What it means |
|---|---|
| Market source | Which exchange’s historical data to use; pick it before the symbols |
| Symbols | The markets to backtest |
| Start date / End date (inclusive) | Calculated in UTC. For a first run, use the 7-day range marked “Recommended”; it loads faster |
| Initial equity (USDT) | Starting equity for the backtest |
| Simulated backtest candle interval | The candle used to replay the price path, 1 minute at the finest. Indicators such as RSI keep the strategy’s own timeframe |
Then check the settings in the “Strategy” and “Risk & execution” tabs. Write down this set of inputs, then change one variable at a time on later runs. See Backtest Settings and Reproducibility for every field and the fill assumptions.
Run and confirm the version was saved
Select Run backtest and wait for it to finish. Once a version exists, the button shows the next version number, such as “Run V2”. After the run, the page saves it as a new version in the experiment. When the version and its parameter snapshot appear under the Saved backtest parameters tab, the save worked.
If you see “Backtest finished, but saving to the server failed”, select Retry save. A local curve alone does not prove the server saved it.
Check what the strategy did
Read net PnL, max drawdown, filled orders, fees, slippage, and final equity, then the trade records, positions, and data warnings. “Closes” only counts reductions and closes, so zero can still mean an open position. Check positions and orders too. Note the experiment name, the version number, and one trade worth inspecting so you can reproduce the result.
After two or more runs, use version comparison to select 2 versions, compare their metrics and configuration differences, and decide which parameters to use.
Use this version for a paper strategy
- Add a paper exchange under Account → Trading account by following First Paper Strategy.
- Return to the experiment, select the version under Saved backtest parameters, choose Use this version’s parameters, and confirm.
- Select Create strategy from parameters. Under “Trading account” in the dialog, choose your paper account, and check the parameters and cost estimate.
- Select Create and start. When you see “Strategy created; start requested”, select View strategy to confirm its runtime status.
The creation dialog uses the parameters currently in the editor; viewing a historical version does not apply them. The dialog also supports live accounts, so confirm your paper account is selected before submitting. Backtest returns do not guarantee later paper or live performance.