Backtesting Basics
How to think about backtesting Turtle-style rules without overfitting or ignoring execution costs.
Backtesting helps evaluate whether a rule set has behaved coherently in historical data. It does not prove that future results will match the test.
What to define first
- market universe;
- entry and exit rules;
- position sizing;
- stop logic;
- pyramiding limits;
- fees, slippage and funding costs;
- data source and cleaning rules.
Avoid common traps
Survivorship bias, look-ahead bias and parameter overfitting can make a system look better than it is. A robust test should include out-of-sample periods, realistic costs and enough trades to make conclusions meaningful.
What to inspect
Do not only look at the equity curve. Review maximum drawdown, longest losing streak, average win, average loss, trade count, market contribution and sensitivity to parameter changes.