Arizona Cardinals vs Los Angeles Rams — Win Probability

Cardinals Stun Rams in 18-16 Upset; Model Massively Underestimated Arizona

Arizona arrived at Los Angeles as a heavy underdog in Week 6, and the model agreed emphatically. The Cardinals carried just a 16% win probability in the pregame projection, while Las Vegas and the model's own assessment pegged the Rams as 13-point favorites. The simulation expected a Rams blowout: LAR 32.3, ARI 19.2, a margin of victory of 13 points in Los Angeles's favor.

The game played out nothing like that. Arizona won 18-16, a 2-point margin that flipped the result entirely and obliterated the expected outcome. That's a 15-point swing from projection to result—the Cardinals exceeded their expected score by essentially nothing, while the Rams fell 16.3 points short of their projection. The model's expected margin of -13 missed by exactly that much in the opposite direction.

This ranks as a decisive model miss, not a close one. The gap between a 16% win probability and an actual victory is substantial, and the magnitude of the scoring miss compounds it. The Rams' collapse from a projected 32.3 points to just 16 carries real analytical weight: either the model overvalued Los Angeles's offensive capacity, undervalued Arizona's defensive pressure, or both. A 16.3-point shortfall for the favored team isn't noise.

For Arizona, the result improved their record to 2-4. For Los Angeles, it left them at 2-4 as well. Both teams left the field at identical records despite the model's conviction that one was vastly superior. The Cardinals' 16% probability converting to a win is the kind of outcome that demands review of the inputs—not as a referendum on the model itself, but as evidence that something in its read on this matchup diverged sharply from reality.