Recency bias is the tendency to overweight recent events in your assessment of what is likely to happen in the future. After three months of market gains, traders expect gains to continue. After three months of losses, they expect losses to continue. The further back in time an event occurred, the less weight it receives in the mental model, regardless of its relevance.
This is particularly damaging at market turning points. At market tops, recent experience has been universally positive, which makes bullish extrapolation feel natural and bearish caution feel unnecessary. At market bottoms, recent experience has been universally negative, which makes bearish extrapolation feel natural and bullish positioning feel reckless. In both cases, recency bias pushes traders toward precisely the wrong conclusion at precisely the wrong time.
In crypto, recency bias is amplified by the speed of cycles. A three-month rally in crypto can produce gains that would take years in equities, creating an intense recency signal that is hard to override. New participants who enter during a rally have no personal experience with drawdowns, so their entire reference set is bullish. This makes the eventual reversal more surprising and more painful for the cohort that entered based on recent performance.
Inflation expectations provide a macro example. After several decades of low and stable inflation, economic models, consumer behavior, and market pricing all assumed inflation would remain low. When inflation spiked in 2021-2022, the initial response was to treat it as transitory, because the recency of decades of low inflation dominated the assessment. The failure to recognize a regime change quickly enough was partly a collective recency bias.
Counteracting recency bias requires deliberately consulting longer historical records. When your instinct tells you that the current trend will continue, look at ten-year charts rather than three-month charts. Look at how similar trends ended in prior cycles. Look at base rates: what percentage of the time does a trend of this magnitude and duration actually continue versus reverse? The long view provides context that the recent view obscures.
Statistical approaches help quantify recency bias. If you are forecasting returns, compare a forecast based only on the last three months of data to one based on the last ten years of data. The difference between these forecasts reveals how much your short-term view diverges from the long-term average. When the divergence is large, recency bias is likely distorting your assessment.
Mean reversion awareness is a natural antidote to recency bias. Most market variables (valuations, volatility, spreads, returns relative to trend) tend to revert to their long-term average over time. Extreme recent readings are more likely to mean-revert than to persist, which is the opposite of what recency bias predicts. Training yourself to think about where a variable sits relative to its long-term average rather than its recent trend helps counteract the bias.
One practical exercise: at the beginning of each month, write down your expectations for the next month, quarter, and year. Note which are based on extrapolating recent trends and which are based on other analysis. After six months, review which forecasts were more accurate: the recency-based extrapolations or the ones grounded in longer-term analysis. Most traders find that their recency-based forecasts are systematically wrong at turning points, which provides the motivation to adjust their process.