Pigskin Nightly
More
7 live
Latest
IND 3FinalHOU 26ARI 7Q2 0:00KC 27Win probKC 99%BAL 26Q2 2:22CAR 22Win probCAR 53%MIA 3Q2 1:24BUF 20Win probBUF 99%NO 7Q2 0:54CHI 16Win probCHI 89%TB 7Q2 0:00DET 10Win probDET 68%TEN 6Q2 1:39DAL 27Win probDAL 97%PIT 3Q2 0:00PHI 13Win probPHI 73%NYJ SunLAC JAX SunNYG LV SunDEN MIN SunSF CIN MonWSH
Falcons DeskThe Verdict

Chiefs earned a 35-13 win in Atlanta, and every snap backs the score

The snaps gave the Falcons a 0% chance against Kansas City, and the Chiefs' deserved margin of 23.1 points sat within a point of the final.

The VerdictDeservedDeserved 0% · Week 10
By Falcons Desk, Pigskin NightlyNov 17, 1:10 PM ET3 min read

The verdict

Played like that, Atlanta wins 0% of the time. The snaps gave the Falcons no chance against the Kansas City Chiefs, and the engine put Kansas City above 99%. The deserved margin was Chiefs by 23.1 points. The real margin was 22, in a 35-13 final.

This one was earned, start to finish. Nothing in the film suggests the scoreboard flattered the winner.

What the score hid

The score hid very little. EPA, or expected points added, measures how much each play moves a team's scoring outlook. The Falcons offense came in at -0.27 per snap, and the Chiefs posted 0.13. Success rate ran 32% for Atlanta and 44% for Kansas City, a 12-point gap.

Luck neither rescued nor robbed anyone. The Falcons finished at -0.4 luck points and the Chiefs at -0.2. The bounces were nearly even and slightly unkind to both sides.

The swing play came in the second quarter with 2:00 on the clock. P.Mahomes threw short right to E.Johnson, who carried it 34 yards to the Atlanta 43 with D.Hellams in coverage. It added 10 percentage points of win probability for Kansas City. That was the biggest single swing of the day, and the gap behind it was built from a long run of snaps that did not work.

The calls

Kevin Stefanski made eight graded fourth-down decisions, and the model disagreed with two of them. Both were punts. Early in the first quarter, on fourth-and-2 from 62 yards out, the model leaned toward going. The punt cost 0.7 expected points, the most of any call in the game. At the top of the second quarter, on fourth-and-4 from 65 yards out, the model called it a tossup that favored going. That punt cost 0.1.

In the third quarter, on fourth-and-4 at the 50, Stefanski did go. The try was stopped, but the model had called it the right decision. The sequence tells the story: he was cautious early, when the game was close enough to take a chance, and he went only after it had slipped away.

The flags split evenly. Each of the four heaviest swung win probability by 4 points, and none came late. Atlanta was hit with defensive offside and a 10-yard offensive holding call in the second quarter. Kansas City gave back a 10-yard holding call and a false start in the first. The officials did not decide this game.

The trend

Atlanta is 3-7, and the snaps say that is about right. The deserved-wins total is 2.8, so luck has added 0.2 wins. That ranks 13th of 32, luckiest first. The Falcons are 0-3 in one-score games, but the luck sheet shows 5.7 total luck points, including 1.1 from fumble recoveries and 2.6 from opponent kicking. Those help a team. They do not hurt one.

The net EPA per snap is -0.118, which ranks 26th of 32. The Chicken Index sits at 67%. The model recommended going on fourth down six times this season, and Atlanta went twice. Across 45 graded fourth downs, the Falcons have lost 7.5 expected points, with one clear miss. They have also been flagged 69 times for 603 yards.

Nothing on the sheet points to a correction by luck alone. The record matches the play, so any change has to come from the snaps themselves.

Next up is the Minnesota Vikings on the road in Week 12, Sunday, Nov 29, at 1:00 PM ET on Fox.

How this was reported. Every number in this piece comes from the play-by-play, drive, injury and snap data Pigskin Nightly collects for every game. EPA (expected points added) is from the open nflverse model. We do not use anonymous sources or invent quotes. Read how we write; spotted an error? Tell us and we will correct it in the open.