Look-ahead bias

Using information in a simulation before it was actually available.

The commonest form is financial-statement data stamped with the period end rather than the filing date. Annual results for a March year end are not public until months later, and a backtest that trades on them in April is trading on information nobody had.

The correction is to lag every fundamental input by a realistic reporting delay, which typically reduces measured performance noticeably.

If the improvement disappears when the lag is added, the strategy was the bias.