Glossary
Backtest
A simulation of a trading rule on past data, to see what it would have done. It gives one path through history, built with hindsight, and is better at rejecting ideas than at proving them.
A backtest runs a rule written in advance over historical prices as if time moved forward, and records the trades, costs and equity curve it would have produced. It is one path through the past, and the designer has usually seen the data first.
Its value is mostly negative: a weak result is a good reason to drop an idea, while a strong one is a reason to test further, with costs, out-of-sample data and a count of the variants tried.
Example: five trades of +2%, -1%, +3%, -2% and +1.5% turn 10,000 into about 10,346, a total return of 3.46%, before costs.
Sources
- Pseudo-Mathematics and Financial Charlatanism: The Effects of Backtest Overfitting on Out-of-Sample Performance. Notices of the American Mathematical Society 61(5), 458-471 (Bailey, Borwein, López de Prado, Zhu), 2014. Accessed 9 Nov 2026