Celta Vigo absolutely dismantled Racing Santander with a 5-0 win that bore no resemblance to the tight affair our model had predicted. Hugo González opened the scoring from the penalty spot in the 32nd minute, then Ilaix Moriba doubled the lead just before half-time with an assist from Miguel Roman. Roman himself got on the scoresheet in the 69th minute, and Pablo Durán Fernández added a brace late on—first in the 78th with setup from Hugo Alvarez Antunez, then again in the 90+2nd from the same provider. It was a clinical, dominant display that left no doubt about the gulf in class on the day.

Our model had leaned toward a Celta Vigo win at 43% probability, so we got the result direction right, but the scoreline was nowhere near the predicted 1-0. The pre-match case rested on Celta's home motivation—sitting in relegation trouble with minimal margin for error—against a Racing side with nothing pressing to play for. That motivation clearly showed up. What didn't materialize was the low-scoring pattern we'd flagged based on Celta's weak home output historically and Racing's travels on the road. Instead, the hosts broke through early and never stopped, turning what looked like a competitive fixture into a rout.

There's no getting around it: we underbacked this scoreline badly. The model had weighted a tight, cagey game; Celta had other ideas and executed a performance that left no room for argument. Getting the winner right counts for something in probability terms, but when the actual margin looks this different from the forecast, it's a clean miss on what the match shape would actually be.