What is an example of look-ahead bias?

Testing a strategy that buys on strong annual earnings dated to the last day of the fiscal year. The 10-K reporting those earnings is filed weeks or months later, so the backtest trades on numbers nobody had yet. The same error appears when a monthly economic figure is placed on the month it describes rather than on the day it was published.

What is the difference between look-ahead bias and survivorship bias?

Look-ahead bias leaks future information into a past decision, such as a number published after the trade date. Survivorship bias leaks future knowledge of which companies would survive, by testing only on today's survivors. Both make a backtest look better than it could have been in real time, and both are prevented by rebuilding the data as it stood on each date.

Are the Opulence Alpha studies free of look-ahead bias?

They are built to avoid it, and the methodology paper sets out how. Chart-pattern detections use only price bars up to their own day, and a turning point counts only once later bars confirm it. But signals are recomputed over the past with today's definitions and data, so a result is a measured tendency, not a live record. And each return starts at the close the signal is computed from, while a real trade would start later.

FAQ · Look-ahead bias

What is look-ahead bias in backtesting?

Reviewed against the platform's code on Sep 26, 2026

Look-ahead bias is the error of letting a backtest use information that was not yet available on the date of the simulated decision — earnings before they were filed, economic data before it was released, or a price that closed after the trade. It makes a strategy look better than any investor at the time could have done.

Why it matters

The leak is usually small and invisible: a quarterly report dated to the end of its quarter although it was filed weeks later, a revised GDP figure standing in for the first estimate, or a model trained on data that overlaps its test period. Each hands the strategy part of the answer, and the backtest cannot tell the difference. Banz and Breen (1986) showed that look-ahead and after-the-fact selection in accounting databases can change measured results.

How it works

Every value is stamped with the moment it became public, and a simulated decision may read only values stamped before it: fundamentals by filing date, economic series by release date, ideally in their first-release vintage (Croushore & Stark, 2001), and prices up to the decision's close. In model validation, walk-forward splits, a purge of overlapping labels and an embargo keep training data from seeing the test period (López de Prado, 2018).

How Opulence Alpha applies it

Opulence Alpha reads company filings from their SEC filing date; its insider-trading and ownership features count a filing only from the next session, because EDGAR can date a filing made after the close with that same day. Economic figures enter only after a per-series publication delay set on the cautious side, but the stored values are the latest revision and the NBER recession flag is dated after the fact: point-in-time in timing, not in vintage. Strategy validation purges and embargoes on the trading-day calendar.

How the studies are measured, limits included →

Related concepts

Point-in-time dataPoint-in-time data records what was actually known on each date — filing dates, revisions, delistings — so a backtest cannot use information from the future. Lookahead and survivorship bias explained.Survivorship biasSurvivorship bias: testing only on companies that still exist. Why it flatters backtests, how to correct it, and how the Opulence Alpha studies handle it.Walk-forward validationWalk-forward validation tests a strategy the way it would have been used: fit on the past only, test on the next unseen period, then roll forward and repeat.

All concepts →Sector behaviour in every market regime →

Questions

What is an example of look-ahead bias?: Testing a strategy that buys on strong annual earnings dated to the last day of the fiscal year. The 10-K reporting those earnings is filed weeks or months later, so the backtest trades on numbers nobody had yet. The same error appears when a monthly economic figure is placed on the month it describes rather than on the day it was published.
What is the difference between look-ahead bias and survivorship bias?: Look-ahead bias leaks future information into a past decision, such as a number published after the trade date. Survivorship bias leaks future knowledge of which companies would survive, by testing only on today's survivors. Both make a backtest look better than it could have been in real time, and both are prevented by rebuilding the data as it stood on each date.
Are the Opulence Alpha studies free of look-ahead bias?: They are built to avoid it, and the methodology paper sets out how. Chart-pattern detections use only price bars up to their own day, and a turning point counts only once later bars confirm it. But signals are recomputed over the past with today's definitions and data, so a result is a measured tendency, not a live record. And each return starts at the close the signal is computed from, while a real trade would start later.

References

Banz, R. W. & Breen, W. J. (1986). Sample-Dependent Results Using Accounting and Market Data: Some Evidence. Journal of Finance 41(4).
Croushore, D. & Stark, T. (2001). A Real-Time Data Set for Macroeconomists. Journal of Econometrics 105(1).
López de Prado, M. (2018). Advances in Financial Machine Learning. Wiley.

Educational content about research methods. Not investment advice.