Minnesota Vikings vs New Orleans Saints — Win Probability

MIN beats NO as model's close call plays out closer than expected

Minnesota's 23-13 win over New Orleans matched the market's pregame read but diverged sharply from what the model projected would happen inside the game itself.

The model assigned Minnesota a 57% win probability before kickoff, making the Vikings a modest favorite—a view that aligned with the market's 2.5-point spread. But the model's expected scoring suggested a much tighter contest: 22.0 for Minnesota, 19.5 for New Orleans, yielding an expected margin of just 2.5 points.

The actual result was a 10-point victory, exactly double the model's projected margin. That gap matters. Minnesota did win as favored, so the directional call was sound. But the model significantly underestimated how much separation there would be.

The Vikings' 23 points came in 1 point above the model's expectation. New Orleans' 13 fell 6.5 points short of the model's projected 19.5—a meaningful miss on the Saints' offensive production. That underperformance drove most of the discrepancy.

In terms of model calibration, this sits in an awkward middle ground. The outcome itself—a Minnesota win—was correctly identified with reasonable confidence (57% was not an overwhelming lean). But the magnitude of that win wasn't captured. A 2.5-point expected margin that actually played out as 10 points suggests the model's read on New Orleans' scoring potential was too optimistic, or on Minnesota's defensive capability was too conservative.

This is not a model failure in terms of picking the winner. It is, however, a meaningful miss on game script. That distinction matters for bettors working with spreads or totals, and for analysts tracking whether the model's internal game projections align with external market pricing over time.