Colts 19, Texans 17: Eight Pundits Picked Houston, One Got It Right

The Colts won this one 19-17, and the pre-game record on it reads almost like a dare: nine graded calls settled by the final whistle, eight of them backing Houston to win, one backing Indianapolis. Indianapolis won.

That's not a split so much as a stampede in the wrong direction. Every score prediction and every outcome pick we tracked for this game had a pundit on record before kickoff, and only one of the nine had the right team raising its hands afterward.

The one call that held up

Tyler Sullivan projected Colts 26, Texans 23 — a three-point Indianapolis win. The final was Colts 19, Texans 17, a two-point Indianapolis win. That's about as close as a score prediction gets without landing exactly, and it graded out at 97.62 accuracy, the best mark on this slate.

Worth being straight about what that call was, though: Indianapolis was the home team, and the site's own naive baseline for this game already leaned toward the Colts by a hair (a 0.47-point projected home margin). Picking the Colts wasn't a road-underdog call against the grain — it was the favored side. That's why Sullivan's skill score for it was a modest +2.39 rather than a big number: correct, and worth crediting, but not a call that beat the crowd's basic assumption so much as it agreed with it and then nailed the number.

The misses

Nobody missed by more than Pete Prisco, who picked "Texans 31, Colts 20", forecasting an 11-point Houston win in a game Indianapolis won by two. That call graded 28.57 accuracy, the week's worst on this board.

Richard Sherman came closer on the shape of the game than most while still getting the winner wrong: he had Houston winning 17-14, a three-point margin — the same margin Sullivan called, just assigned to the other team. It's a reminder that a plausible-looking final score is only half the job.

Jared Dubin had Texans 20, Colts 16, Jordan Dajani had Texans 27, Colts 21, and John Breech was on Houston twice over — 27-23 in a season simulation and 27-24 in his weekly picks. All four graded in the high-40s on accuracy: wrong winner, but not wild on the shape of the game.

The flattest misses belonged to the two pundits who picked a winner with no score attached. Gregg Rosenthal went in leaning on his season-long case for Houston, writing before kickoff that "the next time you see Nick Shook and I, he will be recapping that big Texans win against the Colts. I'm sorry, Colts." That graded zero accuracy. John Granato went further out on the limb back on May 19 — months before the season started — telling Daniel Jones and the Colts "you got no shot against this defense." That call carried the longest lead time of the nine and graded zero as well.

What this does and doesn't say about the pundits involved

None of this looks like a pattern for the people involved — it looks like one bad week inside a much bigger sample. Rosenthal has graded 98 game-outcome picks overall at 56.66% accuracy, so Sunday's miss is one entry in a track record, not a trend. Sherman's career score-prediction accuracy sits at 68.99% across 23 graded calls, well above the 47.62 this one produced — a below-average week for him, not a new normal. Prisco's score-prediction accuracy overall is 62%, more than double what Sunday's 31-20 call managed. And Sullivan's game-score accuracy sits at 59.36% across 22 graded calls — this hit was a good result, but for him, a fairly normal one.

The bigger picture

Across everything we've graded involving these two clubs, calls on the Texans have carried a modest positive skill edge overall (+8.05 average skill delta across 237 graded calls), while calls on the Colts have run essentially break-even (-0.87 across 152 graded calls). Sunday flipped that script for a week: the field's Texans optimism cost most of the pundits above, and the lone Colts believer walked away with the best grade on the board. Anyone still holding a Houston season-win-total or division call should note how quickly that optimism ran into a two-point road loss.

For the season-long picture on both teams, see the Colts and Texans pages, and see who's actually calling games well on the leaderboard.