Backtesting

What does your backtest actually assume?

Inspect data quality, transaction costs and rule behaviour before drawing conclusions from a historical simulation.

Historical test settings and annotated trade events on a monitor

Test the rules and document the gaps

We configure historical tests for a defined strategy using available market data, explicit order assumptions and a recorded version of the rules. The work normally moves from data review to test configuration, repeatable runs and an interpretation session over several weeks. At the end, the client receives test settings, run outputs and a note of material limitations, including data gaps and simulated execution. Historical results are not a reliable prediction of future trading outcomes.

Inputs that make a test easier to interpret

A useful backtest begins with assumptions written down before its charts are reviewed.

  • A dated version of the trading rules
  • The instruments and time periods under review
  • Assumptions for fees and order execution
  • A record of missing or adjusted data
  • A separate period for additional checking

What the review covers

Data provenance

We identify the data used and flag visible gaps that could change how a strategy appears to behave.

Execution assumptions

We record how orders, fees and delays are represented so simulated trades are not mistaken for broker confirmations.

Repeatable runs

We preserve settings and rule versions so a later change can be compared with the earlier test.

A simulation is not a live account

Historical data cannot fully capture future liquidity, unexpected interruptions or every price available when an order reaches a broker. Repeatedly adjusting rules to fit one dataset can also make a test look more persuasive than it is. We present the assumptions alongside the outputs and encourage a separate review of how the strategy behaves outside the period used to develop it. The decision to trade, and the risk that follows, remains with the client.

Bring a rule set for review

Tell us what you want to test and which data or prior results you already have.

Ask about backtesting
Reviewing data-gap notes beside a backtest chart