Test

With your algorithm created, it’s time to see how it would have performed on historical data. Test mode is where you backtest, inspect trades, compare runs, and check whether the strategy generalizes.

Note

Coming from Build? Click Quick Test on a Config node to open Test with that algorithm pre-selected.

1. Select an Algorithm

   Switch to Test from the header mode selector. Open the algorithm picker and choose the strategy you want to test.

   The Test workspace loads the saved strategy and shows its current version, symbol, entry rules, exit rules, protection rules, and any launch blockers.

2. Configure the Backtest

   In the Setup tab, choose the run context:

   Setting	Suggested Start
   Start	6 months ago
   End	Today
   Symbol	BTC/USDC or your market
   Capital	Your intended test budget
   Leverage	Start low
   Trades limit	200

   Caution

   Use at least 3-6 months of data for meaningful results. Short backtests can be misleading — a strategy might look great over 2 weeks but fail over longer periods.

3. Make Local Test Edits

   The Entry, Exit, and Reset tabs let you tune the local test draft:

   • Entry — position size, long/short entry conditions, and session gate behavior
   • Exit — long/short exit conditions, stop behavior, and protection context
   • Reset — cooldown, timeout mode, and follow-up chain behavior

   Note

   Draft edits stay inside Test mode. They do not automatically change the saved strategy in Build. Use this to experiment safely, then make permanent changes in Build once you know what worked.

4. Run the Backtest

   Click Run Backtest. Reversion submits the test, tracks progress, and loads results when the run completes.

   The engine uses the same strategy logic as live execution — entries, exits, stop loss, take profit, trailing logic, fees, slippage, and state transitions are all simulated through the same model.

5. Read the Summary

   Start with the Summary tab:

   Metric	What It Tells You	Good Range
   Win Rate	% of trades that were profitable	> 40% (depends on R:R)
   Profit Factor	Gross profit / gross loss	> 1.5
   Sharpe Ratio	Risk-adjusted return	> 1.0
   Max Drawdown	Largest peak-to-trough decline	< 20%
   Total PnL	Net profit or loss after costs	Positive

   Ask yourself:

   • Sharpe > 1.0? The strategy has risk-adjusted edge.
   • Max drawdown < 20%? The risk is manageable.
   • Win rate + profit factor make sense together? A 35% win rate with 3:1 R:R can be fine.
   • Equity curve smooth? Or does all profit come from a few lucky trades?

6. Inspect Trades

   Open the Trades tab to see every completed trade. Click a row to open Detail, where you can inspect:

   • Entry and exit price
   • P&L and P&L %
   • Exit reason
   • Stop owner and trail mode
   • Position sizing diagnostics
   • Adaptive stop or take-profit resolution, when used

   This is where you find out why a strategy made money or lost money, not just whether it did.

7. Replay the Run

   Use the replay controls under the chart to walk through the backtest over time. The chart and equity curve move together, so you can see where entries, exits, stops, and drawdowns happened.

   Replay is especially useful when the headline metrics look good but the equity curve has sharp drops. Watch the losing periods and decide whether the strategy behavior still makes sense.

8. Compare Runs

   After a completed backtest, switch to Compare. Run another backtest after changing settings, then compare:

   • Current run vs previous run
   • P&L, drawdown, Sharpe, and trade count
   • Exit mix
   • Diagnostics coverage

   Note

   Keep one clean baseline in mind while you experiment. It makes it much easier to see whether a change actually improved the strategy.

9. Run Optimize

   Switch the Test workspace from Backtest to Optimize. Optimize runs rolling cross-validation: it slices the date range into repeated train and test windows so you can see whether the strategy holds up across different periods.

   Configure:

   Setting	What It Does	Suggested Start
   Train Days	Window used to establish behavior	30
   Test Days	Out-of-sample window checked after each train	7
   Step Days	How far the window slides forward between folds	7

   Click Run Optimize. When it finishes, review the aggregate test metrics and the Folds tab.

   Caution

   If train windows look much better than test windows, the strategy may be overfit. Consistent fold performance is what you want before going live.

---

What’s Next?

Want to…	Go to
Run your algorithm live	Run Walkthrough →
Learn all Test metrics in depth	Test Features →
Tune the saved strategy permanently	Build Features →