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
Cowboys DeskThe Verdict

Cowboys lost to the 49ers by six and played like a team that lost by 14

Dallas scored 40 and still had a 1% deserved win probability, because San Francisco gained 0.42 EPA per snap to the Cowboys' 0.13.

The VerdictOffenseDeservedDeserved 1% · Week 10
By Cowboys Desk, Pigskin NightlyNov 17, 9:10 AM ET2 min read

The verdict

Played like that, the Cowboys win 1% of the time. The snaps said 49ers by 14.2 points, and the final was 46-40. San Francisco earned this one.

The scoreboard was kinder to Dallas than the film. The deserved margin was 8.2 points wider than the six-point gap on the board. A 40-point night at home should not read as a beating, but the data says it was.

What the score hid

The Cowboys averaged 0.13 EPA per snap. EPA is expected points added, the points a play is worth over what the down and distance predicted. San Francisco averaged 0.42, a gap of 0.29 on every snap. Dallas was positive, but the 49ers were far better.

Success rate was close: 47% for the Cowboys, 49% for the 49ers. San Francisco's edge came in how big its good plays were, not in how many it had.

Luck did not explain the loss. The Cowboys got 1.1 luck points and the 49ers lost 2.3. The breaks leaned toward Dallas, and the Cowboys still lost.

The swing play was in the second quarter. At 2:43, B.Purdy threw short right to C.McCaffrey, who took it 74 yards for a touchdown. That one snap moved the 49ers' win probability by 21 points, the largest swing of the game.

The calls

Head coach Brian Schottenheimer was mostly in line with the model. Of six graded fourth downs, five matched the best option. The Cowboys kicked on fourth-and-7 from the 34 and again on fourth-and-6 from the 34, and went on fourth-and-1 twice, converting both.

The one split came in the second quarter. On fourth-and-4 from their own 32, the Cowboys punted when the model preferred going. It was a tossup and cost 0.1 expected points, so it is not what lost this game.

The last call was a go on fourth-and-2 at the 40 in the third quarter, with Dallas at 11% win probability. It was stopped. The model agreed with the decision, and the issue was how far behind Dallas already was.

The flags were minor. Roughing the passer on San Francisco in the first quarter gave Dallas 15 yards and a 5% swing. A Dallas false start in the second quarter and a defensive offside in the third each cost 3%. The Cowboys have 79 penalties and 702 penalty yards this season.

The trend

Dallas is 4-6 and the snaps say 4.2 deserved wins, a luck figure of -0.2 that ranks 19th of 32, luckiest first. This record is close to honest. The Cowboys are 2-5 in one-score games and have been robbed in one game. They have no wins while lucky.

Their season net EPA per snap is -0.025, 20th in the league. The season luck total is -4.7 points, with opponent kicking luck at -5.8 and fumble recovery luck at 0.5.

On fourth down, the Chicken Index is 30%, 21st of 32 with the most timid first. Schottenheimer has gone for it 7 times in the 10 spots where the model said go, across 41 graded fourth downs. The cost is 3.1 expected points, with one clear miss.

The next test is the Tennessee Titans at home, Sunday, Nov. 22, at 1:00 PM ET on Fox. Dallas will bring a 0.13 EPA offense against a team that has to defend it in Week 11.

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.