Backtesting

Market vs. simulated backtests, and how to create your first one.

A backtest runs a strategy revision against historical data and reports how it would have performed. QTSurfer has two kinds:

• Market backtests — a strategy run against real historical data from a real exchange. This is the standard, and the one this page walks through.
• Simulated backtests — a more advanced variant built for parameter sweeps: you define a reduced sweep, run it, then submit a refinement to narrow it further inside the same simulated backtest as a new phase. Reach for these once you’re optimizing an existing strategy rather than validating a new one.

Create your first backtest

Open Backtesting → New. The flow is a six-step wizard; each step must be completed before the next unlocks, and you can always go back.

1. Exchange

Pick the exchange whose data and instruments the backtest will use.

2. Data

Choose what market data feeds the simulation and at what granularity (the cadence — from raw ticks up to daily bars; real-time cadence uses every update and ignores any multiplier). Higher resolution is more faithful, at the cost of a slower run.

3. Strategy

Pick which strategy (and revision) to test.

4. Instruments

Pick the counterpart instrument(s) and the historical date range the backtest will run over.

5. Capital & costs

Set the initial capital (in quote currency), the percentage of capital allocated per trade, and the fee assumptions:

• A percentage fee applies to every trade.
• An absolute fee (buy/sell, in quote currency) overrides the percentage when set.

This is also where risk assumptions for the simulation live.

6. Review

Confirm every step’s summary and create the backtest. It’s queued for execution — the Backtests card on the dashboard and the best backtests table are where you’ll see it once it has a result.

Reading a result

Every completed backtest and execution is ranked by Sharpe ratio and total PnL — the same two figures the dashboard’s “best backtests” and “best executions” tables sort by, so a strong run is easy to spot without opening it.

Next step

Once a backtest is running well, two natural directions: share the strategy on the Marketplace, or open Laboratory to iterate on the next idea before writing more code.