San Francisco 49ers vs Seattle Seahawks — Win Probability

49ers Stun Seahawks as Model Whiffs on Seattle Upset

San Francisco's 32-29 victory over Seattle in Week 5 was an upset—and it was the kind that exposes real model error, not just variance.

Before kickoff, Matrix's simulation gave the 49ers only a 41% win probability despite being 5-0. The model projected San Francisco to score 21.7 points and Seattle to score 24.7, forecasting a 3-point Seahawks victory. The market agreed on the direction: Seattle was favored by 3 points.

What actually happened: San Francisco outscored Seattle 32-29, winning by 3 points. That's a 6-point swing from the model's expected margin of -3.

The gap matters. The model didn't just miss the winner—it reversed the outcome entirely while getting the spread magnitude almost right. San Francisco scored 10.3 points above projection. Seattle scored 4.3 points above projection. The 49ers' offensive performance was the primary driver of the upset; the model significantly underestimated their scoring capacity in this matchup.

This wasn't a coin-flip miss. A 41% win probability for an undefeated team on the road suggests the model saw something structural that didn't hold up. Whether that was overweighting Seattle's strength, underrating San Francisco's consistency, or both, the model's read of this game was materially wrong.

The 49ers improved to 5-0. Seattle fell to 3-2. Both teams executed better or worse than pre-game expectations suggested they would. For a projection system calibrated to factual accuracy, this is the kind of miss worth examining: not close, not a reasonable variance band, but a complete reversal of the likely outcome.