Pittsburgh Steelers vs Cleveland Browns — Win Probability

Model Missed Cleveland's Upset Win by Nearly Six Points

The Cleveland Browns pulled off a 27–24 upset over Pittsburgh on Sunday, a result that diverged sharply from what the pregame model projected. Before kickoff, the model gave the Browns only a 41% win probability and expected them to lose by 2.9 points. The market agreed, favoring Pittsburgh by 3 points.

What actually happened was different. Cleveland won by 3, flipping the script entirely. That's a swing of 5.9 points from the model's expected margin to the real result—a meaningful miss that goes beyond the typical noise of a single game.

The model's expected score was Cleveland 17.8, Pittsburgh 20.7. The actual final was 27–24, which means both teams significantly exceeded their projected outputs. Pittsburgh scored 3.3 points more than expected; Cleveland scored 9.2 points more. The Browns' offensive outperformance was particularly pronounced and directly drove the upset.

This was not a narrow miss. The model placed Pittsburgh as the likely winner with moderately high confidence (59% win probability), and the Browns not only won but did so by a larger margin than the expected gap in the other direction. For a pregame projection, this represents a notable forecasting error—the kind that invites review of what the model may have underestimated about Cleveland's capacity or Pittsburgh's vulnerabilities in this matchup.

Both teams improved their records with the result: Cleveland moved to 3–1, Pittsburgh to 2–2. The upset is confirmed, the margin is clear, and the gap between expectation and reality is substantial enough to stand out in the model's performance log for Week 4.