2026-05-15: $POET predicted +3.22%, did -11.02%

On May 15, five stock pattern picks were submitted. Two moved in the predicted direction, three did not. The portfolio averaged a loss of 4.98% across the sample. The largest gain came from FIG, which returned 6.28% against a 0.48% forecast. The largest loss came from POET, which fell 11.02% despite a predicted advance of 3.22%. The designated hero pick was POET, labeled bearish but predicted to rise 0.48%. Instead it declined 11.02%, producing a 14.24 percentage-point error in absolute terms. This represents the worst outcome of the day and the directional miss that most exceeded the magnitude of prediction. The miss was substantial and unambiguous. The methodology's calibration this week shows a directional accuracy rate of 40% on a five-pick sample. Prediction magnitudes have varied—some forecasts landed within reasonable distance (FIG overstated gains by roughly 5.8 percentage points), while others showed errors spanning double digits. The POET result in particular suggests pattern recognition on this name produced a signal misaligned with actual price behavior. Continued observation of prediction accuracy against realized moves will clarify whether these misses reflect noise inherent in single-day forecasting or signal degradation in the underlying approach.