Detroit Lions vs Arizona Cardinals — Win Probability

Detroit's 45-21 Win Over Arizona Far Exceeded Model's Conservative Projection

Detroit's 45-21 victory over Arizona was not an upset—the model favored the Lions at 66% win probability pregame—but the final margin obliterated the projection by a substantial distance.

The model's pregame simulation expected Detroit to score 30.1 points and Arizona 24.5, yielding an expected margin of 5.5 points. The market betting line matched that projection exactly at 5.5 points, Detroit favored. Both the model and the market were aligned on a moderately comfortable Lions win.

What actually happened was fundamentally different in scale. Detroit scored 45 points—15 more than projected—while holding Arizona to 21, a full 3.5 points below expectation. The 24-point actual margin was 18.5 points wider than the model forecast.

This represents a significant miss, though directionally correct. The model called the winner and the favorite accurately, and its 66% confidence in Detroit was vindicated. But the magnitude of Detroit's dominance was not captured in the pregame simulation. Arizona's offensive production fell well short of the model's already modest 24.5-point expectation, while Detroit's offensive output exceeded projection by nearly 50%.

For a model whose entire function is precision in point-spread and score projection, an 18.5-point gap between expected and actual margin is material. The Lions were better than the model assessed, or Arizona was worse, or both. The 1-4 record Arizona carries into week 6 suggests underlying weakness that may not have been fully weighted. Detroit's 3-2 record sits between its pre-game projection strength and post-game reality.

The model got the game right in binary terms. It got the game wrong in magnitude.