The Efficiency-Return Paradigm Trap: Decoupling Predictability from Performance in Technical Trading Rules Across Equity Markets

This study challenges traditional informational efficiency testing frameworks by exposing “The Efficiency-Return Paradigm Trap”, where nominal returns and intrinsic predictability are structurally conflated. Traditional econometric procedures remain highly susceptible to Type I errors. They frequently fail to separate random macro market drift from parameter-driven data-snooping distortions. To resolve this blind spot, we introduce the Monte Carlo random-signaling protocol (SMARS). This protocol serves as a complementary, non-parametric diagnostic grid for technical trading rules. The SMARS algorithm utilises the unoptimised average ensemble of the moving-average family to absorb parameter mining distortions. It generates 5000 drift-neutralised synthetic paths. Crucially, these paths leave the underlying sequential data architecture undisturbed, substantially mitigating directional and distributional return distortions. Applied to historical closing prices across the New York, Tokyo, and Euronext-Athens Stock Exchanges, our framework expands econometric inference into a rigorous Three-Outcome Spectrum. It explicitly identifies regimes that are significantly better, not different, or significantly worse than the market baseline. Empirical outcomes demonstrate that nominal outperformance can expand purely as an artifact of macro-trend momentum. Conversely, genuine forecasting capacity can co-exist with nominal net losses. Consequently, the SMARS protocol delivers a robust, bias-corrected baseline for assessing Weak-Form Market Efficiency.