49ers 36, Cardinals 30: The Field Called the Winner, Whiffed on the Blowout
San Francisco beat Arizona 36-30 in Week 3, and we've graded seven calls the game settled — five exact score predictions and two straight winner picks. On the only question that decides wins and losses, the field was unanimous: every single one of them had the 49ers.
That's a shrug on the win-loss ledger. Nobody backed the Cardinals, so there's no upset call to celebrate and no home-cooking miss to point at. But a 6-0 record on the winner hides a much messier record on the number, and that's where this one gets interesting.
Everybody had San Francisco. Almost nobody had the margin.
The final margin was six points. Here's what the five score-callers actually predicted:
Three of the five — Richie Bradshaw, Richard Sherman, and Pete Prisco — all landed on the same shape of game: a 49ers win by 13. Sherman put it bluntly beforehand: "So I expect him to put up 30 points. I expect him to hold Arizona to 17." Arizona scored 30. Bradshaw (27-14) and Prisco ("49ers take it. Pick: 49ers 30, Cardinals 17") called nearly identical lines. All three graded to 83.33% accuracy and a skill score of -6.61 — worse than the site's naive model, which essentially treated this as a near-coinflip decided by home field (a margin of well under a point) and still scored 86.83. Picking the right winner by a wide margin isn't enough on its own; the model's dumb guess beat three sharp-sounding scorelines here.
Two calls did better. John Breech had it "49ers 24-16 over Cardinals" — a predicted 8-point win, closest of the bunch to the real 6-point margin — and that call graded to 95.24% accuracy with a skill score of +15.89, the best mark of the night. Tyler Sullivan was next, projecting "49ers 33, Cardinals 23," a 10-point margin that graded to 90.48% accuracy and +6.89 skill. Both actually beat the model's naive guess rather than just riding a correct pick to a decent-looking number.
On the plain winner-only side, Colin Cowherd ("I look at this game to be an absolute blowout as well... the Niners are clicking on different cylinders") and Nick Winkler both graded to 100% accuracy — but zero skill, because picking the home favorite here is exactly what the model expects by default. Winkler's specific line is worth flagging anyway: back in September he said, of San Francisco's schedule, "I see a three and o start for the niners," and this result made him right down to the number.
No misses on the winner — but three calls worth remembering
Fairness cuts both ways here. Nobody picked Arizona, so there's nothing to pile on for calling the wrong team. The three 13-point calls from Bradshaw, Sherman and Prisco weren't bad reads of who'd win — they were just generic ones, and the game's actual competitiveness (Arizona scored 30, not 17) punished the round-number confidence.
For Sherman, it's a small blemish on an otherwise solid category: his career score-prediction average sits at 68.99% accuracy and +16.97 skill across 23 graded calls, so this one graded well below his own norm. Sherman also turns up elsewhere in this week's ledger for an unrelated, fully blown winner pick, so it's shaping up as an uneven week for him rather than an isolated slip.
Breech's outlier goes the other way. His score-prediction average is 44.72% accuracy and a -16.33 skill score across 55 graded calls — a rough track record overall — which makes this +15.89 call one of his better nights on record, not a continuation of anything.
Sullivan's positive skill score also tracks with his standing average of +10.88 across 22 graded score calls — a solid category for him, and this one fits right in rather than standing out.
What it means beyond one Sunday
San Francisco moves to 3-0 on the season, and the site's tracked calls have been kind to their bigger-picture outlook: pundits predicting a 49ers playoff berth have hit 91.3% of the time across 23 graded calls, and season win-total calls on the team have landed at 68.3% accuracy. Division-winner picks for San Francisco have been the outlier, hitting just 9.1% of the time across 11 graded calls — a reminder that "good team" and "wins the West" haven't been the same bet this year.
Arizona sits at 1-2, and the sample there is thinner and rougher: the tracked division-winner calls on the Cardinals are 0-for-3, and score predictions naming Arizona specifically have graded to just 2.4% average accuracy across four calls — too few to call a trend, but consistent with a team nobody's found an easy read on. For the season-long picture on both clubs, see the 49ers and Cardinals team pages, and check how this week's callers stack up on the leaderboard.