Indianapolis Colts vs Washington Commanders — Win Probability

Model Overestimated Margin Tightness in Colts' Decisive Win

Indianapolis came to Washington as a 5-point favorite and delivered a decisive 30-13 victory, but the game played out far more lopsided than the pregame projection suggested. The model had forecast a 25.7-20.7 scoreline with a 5-point expected margin, giving the Colts a 65 percent win probability. The market's 5-point line aligned closely with that projection. The actual result delivered a 17-point margin—more than triple the expected gap.

This was not an upset. Indianapolis won as anticipated, and the model correctly identified them as the more likely winner. However, the magnitude of the victory exceeded the simulation by 12 points. The Colts scored 4.3 points above their projected output while Washington fell 7.7 points short of the 20.7 forecast.

The gap between expected and actual reveals a model that captured the direction of the game accurately but significantly underestimated the Colts' dominance. A 5-point projection in a game that wound up 17 points wide represents a meaningful miss on degree, even with the winner correctly identified. This is neither a close call nor a complete failure—it falls between those poles as a substantial underestimate of one team's performance relative to the other.

For Indianapolis, moving to 2-2, the blowout nature of the win suggests execution or personnel advantages the pregame model did not fully account for. Washington's 1-3 record after the game came alongside a performance significantly worse than projection, indicating the gap was driven partly by the Colts' superiority in execution rather than random variance alone.

This game serves as a reminder that win probability and point spreads, while predictive of direction, do not always capture the full scope of performance differential that will actually emerge.