New England Patriots vs Seattle Seahawks — Win Probability

Seattle's 3-Point Win Matched Model Projection With Precision

Seattle defeated New England 13-10 in a game where the model's pregame projection proved remarkably accurate, both in outcome and magnitude.

The model projected Seattle to win with a 59% probability and expected a margin of 2.9 points. The market line set Seattle as a 3-point favorite. Seattle won by exactly 3 points, making this neither an upset nor a surprise—a straightforward favorite victory that aligned with both the model's simulation and the betting market's assessment.

The actual final score of 13-10 differed from the model's expected score of 23.7 for Seattle and 20.8 for New England. Both teams scored substantially fewer points than projected, suggesting a defensive-heavy game or offensive inefficiency that the model's pregame simulation did not anticipate. However, this gap in total scoring did not affect the outcome: Seattle still won by the margin the model expected.

This distinction matters for evaluating model performance. The projection of a 2.9-point margin proved accurate to within 0.1 points despite a 10-point swing in combined scoring. The win probability assessment of 59% also reflected a competitive but decidedly favored matchup, which the actual result confirmed.

The model called the winner correctly and the margin correctly, even though the path to that result—a lower-scoring game than anticipated—diverged from the point projections. This demonstrates that margin prediction can hold even when absolute scoring does not. For a model whose value lies in outcome forecasting, this game represents successful calibration: the favorite won as favored, by the amount expected, in a game the model assessed as competitive but not close.