Retail Trader's Ruin: An Anatomy of Popular Signal Failure

Battle Count: 69

Directly addresses whether popular retail trading signals (trend, oscillator, volume, calendar, candlestick) deliver economically meaningful, net-of-cost, survivable edge. The joint falsification framework (statistical + economic + survival gates) is a rigorous template for evaluating any trading strategy. The momentum calibration benchmark design and one-sided claim-exclusion methodology are transferable to institutional strategy evaluation. The paper's conclusion that four of five retail families are REFUTED and none SUPPORTED has direct implications for retail trading strategy development, risk management, and regulatory policy. The leverage/survival analysis under FINRA and ESMA regimes is directly applicable to retail CFD and margin trading.

The framework requires: (1) stationary bootstrap implementation for confidence intervals; (2) hierarchical Benjamini-Yekutieli FDR control across multiple testing layers; (3) exposure-matched benchmark construction; (4) Monte Carlo leverage/margin-call simulation with GARCH-consistent volatility; (5) one-sided claim-exclusion tests; (6) point-in-time constituent membership with Shumway delisting corrections; (7) sensitivity grids across materiality thresholds and cost levels. While each component is established, the joint composition and correct implementation of all three gates simultaneously requires careful statistical programming. The paper provides exact parameters and a replication code promise, reducing practical barriers.

The paper states that source code, experiment manifests, and redistributable derived artifacts will be made available at a permanent public archival location upon publication. Licensed market data are not redistributed but providers, fields, transformations, and audit-mode artifacts are documented. However, there is no pristine holdout (data and rules examined in prior work), no external peer review has been conducted, and the paper is dated July 2026 (very recent). The stationary bootstrap is not a held-out test. Reproducibility is moderate: methodology is fully specified with exact parameters, but data access is restricted and the repository is not yet publicly available.