Full time: Mallorca 2, Real Madrid 1. Time to hold the predictions accountable. **What the numbers said:** Real Madrid were favored at around 45% win probability, with Mallorca given the lowest chance at roughly 29%. The odds had Real at 3.40, Mallorca at 2.375, and the draw at 2.875. The market leaned toward Real or a draw. A Mallorca win was considered the least likely outcome. **What actually happened:** Mallorca took control through Morlanes in the 41st minute and doubled their lead in the second half. Real Madrid looked disjointed, cycling through five substitutions — including Guler (72'), Diaz (76'), and Luvumbo (81') — without finding rhythm. Militao's 88th-minute header was a consolation, nothing more. Mastantuono picked up a yellow in stoppage time to cap a frustrating night. **Where the prediction failed:** The model overweighted Real Madrid's squad strength and underweighted Mallorca's home form and defensive structure. Mallorca at 2.375 was genuinely good value in hindsight. The prediction also missed the possibility that Real's depth wouldn't translate into second-half dominance despite aggressive substitutions. **What it got partially right:** The under 3.5 goals lean was correct — only three goals in total. But getting the winner wrong matters more than getting the goal count right. **Honest verdict:** This was a clear miss. When a model prices the actual winner as the least likely outcome, that is a failure worth noting. Mallorca deserved this result and the data did not see it coming. Credit where it is due — upset confirmed, prediction busted.