Dallas Cowboys vs Green Bay Packers — Win Probability

GB Stuns DAL as Model Overestimated Cowboys' Strength

Green Bay's 23-13 victory over Dallas was a clear upset by the pregame numbers, and the gap between projection and result reveals how the model read this matchup—and where it missed.

Before kickoff, the model gave Green Bay a 47% win probability, making Dallas the favorite despite a modest spread. The model projected Dallas to outscore Green Bay by 1 point (27.3 to 26.3), while the market favored the Cowboys by 1 point as well. Those were closely aligned forecasts: the model and the market agreed Dallas was favored, just barely.

What actually happened was substantially different. Green Bay won by 10 points, posting 23 while Dallas managed only 13. That's a 10-point swing from the model's expected 1-point Dallas advantage.

The upset designation is mathematically clear: a team with only 47% pregame win probability won the game. But the scoring gap is where the real miss emerges. The model projected Dallas would win 27-26 in a close, high-scoring affair. The actual game was a low-scoring defensive battle that Green Bay controlled. Dallas scored 14 fewer points than projected; Green Bay scored 3 fewer.

This wasn't a nail-biter the model called correctly. The model expected Dallas to be slightly better offensively and expected a tighter game overall. The Cowboys' underperformance (down 14 from the 27.3 projection) drove most of the discrepancy. Green Bay's underperformance was minimal by comparison.

The upset itself wasn't shocking given the 47% probability Green Bay received—nearly a coin flip. But the margin of victory exceeding the expected margin by 9 points suggests the model overestimated Dallas's offensive capability or underestimated Green Bay's defensive execution in this specific matchup.