Look-ahead bias: when a backtest accidentally sees the future

The idea in plain English

Look-ahead bias occurs when a test uses information that would not have been available when the simulated decision was made. It can appear in obvious ways, such as tomorrow’s price, or subtle ways, such as revised data, full-period rankings or a closing value used before the close was known.

The distinction that matters

Causal testing requires a clear timeline. For every input, ask when it became available, not merely what date is printed on it. Signals, order submission and fills must follow that timeline. Avoid using future outcomes to choose historical assets or filter trades.

A practical example

A fictional strategy chooses each month’s best-performing asset at the start of that same month. The choice uses the outcome it is supposed to predict. Replacing it with information from the prior period changes the test into something a real decision maker could have attempted.

Try this without moving money

• Draw input availability, signal creation and execution in time order.
• Check for revised data and selection based on future performance.
• Run a version with conservative delays to test sensitivity.

A mistake to avoid

Do not assume shifting one column fixes every timing error. Data selection and preprocessing can leak information too.

Before you act

Every numerical scenario here is illustrative, not a market quote or forecast. A test can be reproducible and still be wrong about the future. SuperQuant describes a research-only workflow without live order execution. Practice in a paper setting, disclose assumptions, and do not confuse a polished chart with a suitable investment.