Commanders 33, Seahawks 31 — Week 3: Twenty Points Off Three Turnovers

Third and nine at the Washington 34, one 55 left, Washington up 33 to 31 and Seattle holding all three of its timeouts. Rachaad White went around right end for 19 yards to the Seattle 47 — eight and a half points of win probability, the model's read on each team's chance of winning at that moment, and the only Washington run of the day to gain 10 yards or more. Seattle spent its timeouts. On third and seven at the Seattle 44 with one 36 left, Marcus Mariota ran around right end for nine, and Washington knelt. The Commanders beat the Seahawks 33 to 31 at home for their first win of the season, never trailed, and by the win probability model never dropped below 51 percent. They did it with 260 total yards, the fifth-fewest of the 28 teams that have played so far this week, and the sixth-most points.

The difference is the takeaways. Three forced, none given away, a turnover margin of +3, tied for third of the 28 teams this week — after no takeaways at all in Weeks 1 and 2. 20 points came off them. First the fumble: 11 oh six left in the first quarter, nothing to nothing, Amik Robertson forced a fumble by Jadarian Price and Tim Settle recovered at the Seattle 21; four plays later Mariota found White for a six-yard touchdown. Then Sonny Styles's interception at the Seattle 42 with 12 10 left in the third quarter, up 17 to 10. Three plays later, second and eight at the Seattle 30, Mariota threw deep left to Terry McLaurin for a 30-yard touchdown. By expected points added — what a snap was worth given the down, the distance and the field position, against what that situation was already worth — that throw was +3.3, and it made it 24 to 10. Then Kain Medrano, with four oh nine left and Washington up 27 to 24: Sam Darnold threw short middle on second and 15 from the Washington 49, Medrano intercepted at the 50 and returned it 50 yards for a touchdown. 29 points of win probability, the biggest swing of the game. Drew Stevens missed the extra point wide left, and it was 33 to 24.

The run defense leads the units, because it is where Sunday and last season disagree most. Seattle ran 18 times for 44 yards, the second-fewest rushing yards allowed by any of the 28 defenses that have played this week, behind only Atlanta's 17. −8 expected points — on defense a big negative number is the good direction — and per play, the best run defense of any team so far this week. 11 of the 18 gained three yards or fewer, and the only one that gained 10 or more was the first snap of the game. At the rate this defense allowed over the first two weeks those 18 plays would have been worth about even; at last season's rate, when it ranked 20-eighth, about +1.

The pass defense gave up the yards and took the ball. 393 passing yards is the most allowed by any of the 28 defenses this week, and the most in any of the 92 team-games this season. Four touchdowns on pass plays and no sacks. But by expected points, 46 Seattle pass plays were worth +5. At the rate this pass defense allowed over the first two weeks, those same plays would have been worth about +12; at last season's rate, when it ranked thirtieth, about +9. The two interceptions are most of the difference. The unit also held at the goal line: first and goal from the five late in the first half, Washington up 14 to seven, Seattle got to the two and then threw three straight incompletions, the last on fourth and two with one 51 left. −3.1 expected points for Seattle, eight points of win probability.

The passing game opens on value against its own baseline, because it beat both. 33 pass plays — a count that includes two sacks — 176 net yards, three touchdowns, no interceptions, +3.9 expected points, and 42 percent of them successful plays, meaning they gained what the situation asked for. At the rate Washington passed over the first two weeks, those plays would have been worth about +3. At last season's rate, when it ranked in the bottom half of the league, about zero. The go-ahead drive came with the score 24 all. The drive started at the Washington 40 because Seattle's kickoff landed short of the landing zone; on third and 13 from the 37, the pass fell incomplete, but AJ Finley was flagged for unnecessary roughness, 15 yards and a first down, 10 points of win probability. Then fourth and one at the Seattle 39, eight 40 left: Mariota short left to McLaurin for five, +2.2 expected points and 11 points of win probability. The drive ran a little over seven minutes and Stevens kicked a 36-yard field goal to make it 27 to 24. Stevens had also matched Seattle's 57-yard field goal with one of his own with three seconds left in the first half, after a drive of 49 seconds.

The run game opens on success rate, and it is the low point: 20 percent of Washington's 30 rush plays were successful plays, the lowest of the 28 offenses this week and third-lowest of the 92 team-games this season. 84 yards, −8.9 expected points, the worst total of any team so far this week, and 23 of the 30 gained three yards or fewer. At the rate of the first two weeks — seventh in the league in Week 2 — those 30 plays would have been worth about +1.7. At last season's rate, when this run game ranked ninth, about plus half a point. The one that got away: second and eight at the Seattle 48 in the first quarter, up seven to nothing, Jacory Croskey-Merritt went 48 yards for a touchdown, and it came back on an illegal formation penalty.

That run game shaped the drive chart. Three of 13 on third down, 20-sixth of the 28 teams this week, and one of the first 11 before the last possession. Five three-and-outs; of the teams that have played this week, only Minnesota had more. 13 drives: three touchdowns, two field goals, seven punts and the kneel-downs. Of Washington's 27 points on offense, 14 came on drives that started in Seattle territory.

For anyone new to this show, here is where that sits. Last season Washington went five and 12 with a passing game in the bottom half by expected points per play, a run game ninth, a pass defense thirtieth, a run defense 20-eighth and a third-down defense in the bottom third. Week 1 was a 24 to 22 loss at Philadelphia, with a run defense that ranked 20-fifth. Week 2 was a 37 to 20 loss at Dallas: the pass defense ranked 20-ninth, the 37 points allowed were the second-most in the league that week, and the third-down defense was thirtieth, while the run game ranked seventh. Sunday turned most of that over. The run defense had the best per-play rate of the 28 teams that have played this week, the pass defense gave up yards but less value than either of its baselines while the defense took the ball away for the first time this season, and the run game went from seventh in the league by the per-play rate to 20-sixth of the 28 so far this week. Washington is one and two. Three games do not settle much, and a win built on three takeaways is hard to schedule. But a defense that ranked in the bottom five against both the pass and the run last season forced every one of them.

Washington faces the Colts in London next Sunday, October fourth.

That's your Commanders — every game, by the numbers. This is Muffed: we do the same deep dive for all 32 teams. Follow the Commanders show wherever you're listening, so next week's recap lands the moment it's out. No theater. Just the football.

And if you play fantasy, there's a Muffed show built around your exact roster — free at muffed.ai.

_Data via nflverse._