Cincinnati Bengals vs Pittsburgh Steelers — Win Probability

Steelers Stun Bengals in Upset; Model Reversed Field Correctly

Pittsburgh entered Week 3 at Cincinnati as a 3-point underdog in both the betting market and the model's simulation. The Steelers' pregame win probability was 41%, with an expected final score of 19.7 to Cincinnati's 22.8 — a projected 3.1-point loss that aligned closely with the market's 3-point spread favoring the Bengals.

What happened instead: Pittsburgh won 30-27, a 3-point victory that stands as an upset by definition. The actual margin was 6.1 points better than the model's pregame expectation for the Steelers.

This was not a narrow miss. The model projected Cincinnati to win and outscore Pittsburgh by 3 points. The Steelers did the opposite — they won and outscored Cincinnati by 3. That reversal amounts to a full 6-point swing from projection to reality.

The scale of the miss matters here. Pittsburgh didn't squeak out a one-point win; it beat a favored opponent by the exact same margin the model had forecasted in Cincinnati's favor. In that sense, the model's confidence in Cincinnati wasn't slightly misplaced — it was fundamentally wrong about which team would prevail and by how much.

Both teams moved to 2-1 after the game. For the Steelers, the upset validates that the pregame 41% win probability, while minority odds, captured real possibility. For the model, the gap between projection and outcome was substantial enough to register as a clear directional error, not a variance within the typical range. Cincinnati was favored for reasons the model agreed with before kickoff. The Steelers' performance on the field proved those reasons insufficient.