Giants 12, Titans 7: The Field Split Evenly on Tennessee — and Every One of Them Lost
Tennessee at New York Giants settled eight graded calls in our Week 3 record, and the field wasn't fooled by consensus — it was split right down the middle, four picks for each side. The Giants won it 12-7. Every single caller who took the Titans lost. Every single caller who took the Giants won.
That's not a unanimous call the field can shrug off, and it's not a unanimous miss either. It's a coin flip that landed on one side, and the four who called tails are feeling it.
The Titans side: nobody survived it
John Middlekauff went with Tennessee on the strength of the quarterback change: "If Jameis is taking snaps, I gotta take the Titans... I like the Titans." That call graded to zero accuracy and a -189 skill score — picking against the home team and losing is the worst version of this bet, not just a wrong guess.
On Colin Cowherd's show, a co-host identified only as "Mitchell" made the same wager: "Give me the Titans... I got Robert Sala in the Titans." Same result: 0 accuracy, -189 skill.
Lane Mills had this loss locked in before the season even started. Back on August 18, his full-season slate for Tennessee listed a "win... over the Giants in game three." That preseason call drew the steepest penalty on the board, a flat -200, the maximum the site hands out for a confident miss.
Tyler Sullivan tried to split the difference with a score, projecting "Titans 23, Giants 20." The call had the winner wrong and the margin backwards — the Giants actually won by five, not lost by three — for 45.24 accuracy and a -83.11 skill score.
The Giants side: right, and two of them earned it
Picking the Giants here wasn't insight — New York was the home team, so backing them was also the site's naive default. Richard Sherman ("I think they're going to have a good day... I got the Giants," graded here) and Tom Pelissero ("These are games you should be able to win, even with a backup quarterback here," graded here) both landed at 100 accuracy — and 0 skill delta, because matching the home-team default earns no credit above the baseline.
The calls that actually beat that baseline were the score predictions. John Breech wrote simply, "Giants 20-16 over Titans" — a four-point margin, one point off the real five-point spread. That call scored 97.62 accuracy and a +15.89 skill delta, the best mark this game produced. Pete Prisco had the Giants by nine ("Pick: Giants 23, Titans 14," graded here) — wider than the actual margin but still on the right side of the ball, worth 90.48 accuracy and a modest +2.39.
What this did to their records
Breech's call is worth a second look because of where he usually sits. Across his tracked score predictions, he's at 44.72 accuracy with a -16.33 average skill score — below his own field average for that call type. This one game outperformed his typical line by a wide margin rather than extending a hot streak.
Middlekauff's miss lands differently. His game-by-game picks are an established sample (22 graded, 67.42 accuracy, +23.29 skill on average) — usually a strength for him. This wasn't a first-time stumble from someone still finding his footing; it's a real subtraction from a category he's normally good at.
Prisco's call was closer to business as usual: across 55 graded score predictions he sits at 62 percent accuracy and a +14.74 average skill score, so backing the Giants by nine was right in his typical range, not a standout or an outlier.
The bigger picture
Season-long, our record on these two clubs isn't close: the Giants own a 71.4 average accuracy across 269 graded calls, while the Titans sit at 63.9 across 177. This result won't move either number much on its own, but it's one more data point on the side of the team that's been easier to call correctly all year. Full team-by-team breakdowns are on the Titans and Giants pages, and the leaderboard has the season-long standings for everyone named above.