Buccaneers 24, Cowboys 16: One in Nineteen Called the Upset

Tampa Bay went into Jerry World 0-4, on a short week, with a backup quarterback. Dallas was favored by just about everyone who said a word about the game. The Buccaneers won 24-16, and in doing so settled 27 graded calls — winner picks, score predictions, and the preview chatter that came with both.

Counted the way the game room counts it — one voice per pundit, their final call standing — 19 pundits picked a side. Eighteen took the Cowboys. One took the Buccaneers. The field's average accuracy on the calls this game graded: 18 out of 100.

That split is about as lopsided as this series sees, and the lean wasn't based on nothing: Tampa Bay was winless, starting Jalon Daniels, and playing on three days' rest. Rich Eisen called it "a backup quarterback in their house" and penciled in a Dallas win that "should destroy Tampa." Caleb Skinner wrote that "it doesn't seem like the Bucs will have much of a chance in this one," pointing to the 0-4 start and a depleted secondary. Colin Cowherd said simply that "Dallas at home should polish off Tampa Bay." All three picked the Cowboys. All three were wrong, and because the Cowboys were also the game's naive default pick, none of them picked up any skill credit for being on the popular side — a plain outcome call on the favorite earns nothing extra for being right, and here it earned nothing at all for being wrong.

The score calls, and how far off they were

Seven pundits put a number on it, and the dots tell the story better than the record book does.

Every single score call had Dallas winning, with predicted Cowboys margins ranging from 10 to 21 points. The Buccaneers won by 8. John Breech and Jordan Dajani came closest, both projecting a 10-point Dallas win (27-17 and 30-20, respectively) — missed by 18 points apiece, the best of a bad set. Walt was furthest off, calling Cowboys 38-17, a 21-point swing from what actually happened.

The single worst grade of the night belonged to Chuck Pagano, who didn't even commit to a final score but said it would "take a miracle" for Dallas not to put up 31 points. It wasn't a miracle; Dallas scored 16. That take graded out worse than any specific score prediction in the group, and it's not an outlier for him — Pagano's overall record on the calls we've graded this season is 0 for 9.

The one call that landed

Clarence Hill picked Tampa Bay before Week 1 even kicked off, calling this "one of those trap games" Dallas could lose. He named it on September 3, more than a month out, and it held up: Hill was the only pundit of the 19 tracked for this game who took the Buccaneers, and because Dallas was the field's default pick, his call beat the naive baseline outright rather than just beating a coin flip.

It's worth saying plainly what that call was worth and what it wasn't. Hill's overall record remains a mixed one — 51.67% accuracy and a negative overall score across a small, still-developing sample of 11 graded calls. This was a real hit on a correct, contrarian, long-lead-time read, not evidence of a hot hand. Both things are true at once.

What this does to the rest of the ledger

For the pundits who missed, the damage varies. Pete Prisco picked Cowboys 34-14 (twice, in separate pieces), missing by 28 points — a bad night, but one bad night inside a much larger, still slightly positive season-long record (52.99% accuracy, +9.34 score across 84 graded calls). Jared Dubin also had Dallas winning big, in line with a season where his score predictions specifically have run below average (43.72% accuracy) even though his outright winner picks have been solid. Walt's miss landed harder in relative terms — his tracked record is only four graded calls, and this one dropped his season score further into negative territory.

The bigger picture

Those numbers line up with a rougher season-long trend for calls about both clubs. Among graded outcome picks involving the Buccaneers this year, the field has been right just 5.9% of the time — pundits have doubted Tampa Bay all season, and mostly correctly, which makes Thursday's result the exception rather than the rule. Outcome picks involving the Cowboys haven't fared much better overall, sitting at 22.6% accuracy across a much larger sample. One upset doesn't rewrite either trend, but it's a reminder that the favorite isn't automatic just because the room agrees on it.

For the season-long view on both clubs, see the Buccaneers and Cowboys team pages, and check the leaderboard for how this week moved the overall standings.