lookback

Lookback is a general term used to describe examining past data or events over a defined period, or the use of past information in determining a result. It is applied across fields such as finance, statistics, medicine, and technology, often to measure history, assess performance, or define a retrospective horizon.

In finance, lookback concepts appear most prominently in two forms. Lookback options are exotic securities whose payoff depends on the maximum or minimum price of the underlying asset during the option’s life, making them highly dependent on the price path. Variants include lookback calls and lookback puts, which can be priced with fixed or floating lookback features. Separately, the term lookback period denotes the window used to measure performance, risk, or exposure, such as calculating returns, volatility, or other indicators over a rolling interval.

In statistics and time-series analysis, a lookback window refers to the amount of past data used to compute a statistic at a given point in time. This approach underpins moving averages, lag features for models, and retrospective forecasting. The chosen window length affects responsiveness to new information versus stability of estimates.

In medical research and clinical practice, a lookback period is the span of time reviewed to identify prior diagnoses, treatments, or exposures, often for study enrollment or outcome assessment. Lookback bias can occur in retrospective studies or backtesting when information from the future inadvertently informs past decisions, potentially distorting results.

In data processing and analytics, lookback concepts support rolling computations, backfill in event streams, and retrospective auditing to evaluate historical performance or behavior.