Base Momentum Fallacy

Momentum fallacy is when people think a trend will keep going just because it already started. They ignore signs that the trend might stop or reverse.

The momentum fallacy describes the cognitive bias of projecting recent directional trends into the future without properly accounting for changing underlying conditions. It manifests in decision processes that overweight recent evidence and underweight structural shifts or mean-reverting dynamics.

A casual investor notices a stock has risen every day for two weeks and assumes it will keep climbing, so they buy in heavily—ignoring that the company's fundamentals haven't changed and that similar short rallies in the past have quickly reversed.

A quantitative analyst builds a momentum strategy that overweights the trailing 20-day return signal and underweights the 12-month mean-reversion prior. During a regime shift—e.g., a sudden tightening of monetary policy—the short context buffer dominates the posterior, causing the model to maintain long exposure well past the inflection point. The resulting drawdown reflects a classic transient state dominance failure: the recency-weighted evidence accumulation framework never triggered the reversal test, because contradictory macro signals fell outside the model's effective temporal window.

Seeing a streak makes someone expect it to continue, so they act on that belief. New opposing signals are downplayed, so the streak looks stronger than it is.

A recency-weighted integration mechanism, anchored to a short-term context buffer, amplifies recent observations relative to historical baseline constraints. The resulting asymmetry in evidence weighting produces a biased posterior favoring continuation over reversion.

Pause and check older information before acting on a trend. Look for signs that the trend might be temporary.

Apply temporal regularization to evidence weighting and enforce horizon-aware priors to reduce recency bias. Insert explicit reversal tests to stress-test continuation assumptions.

Overcommitment to transient signals; Misallocation of resources; Delayed reaction to reversal

Adversarial actors can manufacture or amplify short-term trend signals—through coordinated buying, narrative repetition, or staged data releases—to induce momentum-biased decision-makers to pile into a position or policy direction, then exit before the inevitable reversal. In financial or information warfare contexts, a fabricated streak of confirming evidence can be engineered just long enough to exploit the target's short context buffer before the manipulated signal is withdrawn. Political or marketing campaigns can similarly sustain a manufactured "wave" narrative to suppress defection and recruit fence-sitters who assume continuation.

Practitioners should enforce explicit horizon-aware priors by anchoring forecasts to base-rate distributions and mean-reversion statistics rather than raw recent trajectories. Instituting structured reversal-stress tests—formally asking "what would need to be true for this trend to stop?" before any commitment—directly counters asymmetric evidence integration. Temporal regularization techniques, such as exponential decay windowing balanced against a long-horizon baseline, can be embedded in decision frameworks to mechanically counteract recency-weighting skew.