2026-05-19: $MT predicted -0.26%, did +6.74%

On May 19, twenty pattern picks were tested against actual market movement. Directional calls landed 11 out of 20 correct—a 55% hit rate. The portfolio averaged a positive 0.82% move across all picks, though individual results ranged widely from a 6.74% gain to a 4.05% loss. The hero call was MT (ArcelorMittal), predicted to fall 0.26% but instead rose 6.74%—a 7 percentage point miss on the downside. The bearish signal failed to materialize, yet the stock moved favorably anyway. This represents the largest single-day outperformance in the test set, though it arrived opposite to the predicted direction. CME (Chicago Mercantile Exchange) was the inverse story: a bullish call for +0.75% turned into a 4.05% decline, marking the worst performer. The 55% directional accuracy sits at the boundary of random chance with a twenty-pick sample. The presence of a large positive move that contradicted its bearish forecast—coupled with a correspondingly large miss on the bullish side—suggests the pattern logic this week captured general market turbulence without reliable directional bias. The positive 0.82% average return masks high variance; mean return above zero does not validate the underlying signal. This data warrants scrutiny of whether the patterns are identifying true inflection points or whether recent market conditions simply favored long positions regardless of setup. No claims about this week's efficacy should extend beyond what these numbers show: moderate accuracy, wide dispersion, and at least two outsized reversals that merit investigation into their common features.