The Extremity Premium: Sentiment Regimes and Adverse Selection in Cryptocurrency Markets

Abstract Using the Crypto Fear & Greed Index and Bitcoin daily data (Feb 2018–Jan 2026, N=2,896), we document that sentiment extremity predicts excess uncertainty beyond realized volatility. Extreme fear and extreme greed regimes exhibit significantly higher spreads than neutral periods—the “extremity premium.” Key Findings Extremity premium confirmed: Within-volatility-quintile Cohen's d = 0.21, pooled d = 0.40 (p < 0.001) Strong Granger causality: Uncertainty → spreads (F = 211) Cross-asset replication: Effect replicates on Ethereum and across 6/7 market cycles ABM reproduces pattern: Agent-based model qualitatively matches empirical findings Intensity over direction: How extreme the sentiment is matters more than whether it's fear or greed Status With Editor at Computational Economics (Springer). Working Paper DAI-2510. Links ASCRI: systems.ac/3/DAI-2510 Research Lab: Dissensus AI v3.0.0 (2026-07-11): Matches arXiv v3 (52pp). Central within-quintile deterioration claim relabeled as post-hoc/exploratory (abstract carries the qualifier; momentum control renders it marginal); an SMM J-statistic that traced to no committed code was removed; Table 18/19 count columns corrected to the actual regression samples (N = 719 with warm-up footnote; complete-case note); Table 6 relabeled truthfully as HC3 with a cross-reference disclosing the HAC p = 0.063 on extreme fear; Eq. 15 stated in the linear form implemented in all code paths; Abdi–Ranaldo estimator relabeled as the close-to-close variant; Corwin–Schultz β summation restored in the printed formula; sensitivity-grid recomposition disclosure added.