Backtesting
Backtesting is applying a set of trading rules to historical price data to estimate how the strategy would have performed, in order to assess it before risking capital.
Done properly, it requires rules specific enough to apply without judgement. If a rule needs interpretation, the test measures the interpreter rather than the rule — and the interpreter already knows what happened next.
The failure modes worth knowing
- Hindsight bias — unconsciously applying rules more generously where you remember the outcome.
- Overfitting — adding conditions until the strategy fits historical data perfectly and generalises to nothing.
- Ignoring costs — omitting spread, commission, swap and slippage, which can invert a marginal result entirely.
- Survivorship and sample bias — testing only a period or instrument that happened to suit the strategy.
A backtest establishes that a strategy was not obviously broken historically. It does not establish that it will work, and it says nothing about whether you will follow it under pressure — which is usually the binding constraint.
Related terms
Maintained by Jared Sinclair, Founder · Syrax Global FZCO · Definitions are educational, not financial advice.
Knowing the vocabulary is the easy part.
Every term here can be learned in an afternoon. Applying them consistently under pressure is the part that decides outcomes — and the part we measure.