Dolphins 10, Vikings 15 — Week 4: Four Sacks, Seventeen Throws

Third and five at the Miami 36, Miami's first possession of the second half, Minnesota up nine to nothing. Ollie Gordon took it off left tackle and was not pushed out of bounds until the Minnesota 11, 53 yards later. By expected points added — what a snap was worth given the down, the distance and the field position, measured against what that situation was already worth — +4.7. By win probability, the model's read on each team's chance of winning at that moment, 11 points, the largest swing of the game. Then two yards on three snaps, and a 27-yard field goal. Miami lost 15 to 10 in Minnesota and is oh and four. It ran 37 pass and run plays to Minnesota's 69, and after five 48 of the fourth quarter it never had the ball again.

Start with the pass defense, because it changed. Over the first three weeks Miami allowed +0.48 expected points a pass play; on defense a negative number is the good direction, and that one was the wrong way by a distance. Sunday, 41 Minnesota pass plays, sacks included, came to −0.7 in total, −0.02 a play. 247 yards, no touchdowns, 39 percent successful, meaning they gained what the situation asked for. And four sacks, the first four of Miami's season after none in three games: Sieler on the first snap of the game and again inside the last minute of the half, Uche on third and 11, Robinson on third and one. That last one: two 43 left in the third, Minnesota up 12 to three at its own 35, needing a yard, and Robinson brought Kyler Murray down for a loss of 11. −3.5 expected points for the Vikings, seven points of win probability, and a punt. They came in a game Miami mostly rushed four or fewer: five or more pass rushers on 10 of 43 dropbacks, 23 percent, 20-third of the 30 charted defenses and below its own 32 percent over the first three weeks. What did not change was where the ball went. Murray was catchable on 84 percent of his throws, level with the 84 this defense allowed before Sunday, and contested on one of 37. 13 of his 27 completions went to T.J. Hockenson, who caught every target.

The run defense opened on a light box and gave a little. Miami showed six or fewer defenders in the box on 16 of Minnesota's 26 designed runs, 62 percent, against 29 percent over the first three weeks. Those designed runs gained 4.3 a run, against 3.6 allowed before Sunday. By expected points it held about level: 28 rush plays, a hundred 28 yards, −2.6, −0.09 a play against a season rate of −0.13. Aaron Jones carried 20 times for 94. The stand that mattered was near the goal line: Minnesota ran eight plays inside the Miami 20 and scored no touchdowns, and the defense stopped nine of 17 third downs. 15 points came on five field goals.

The passing offense was the worst unit on the field. 19 pass plays, −12.4 expected points, −0.65 a play, against −0.09 over the first three weeks. Malik Willis went nine of 17 for 85 yards, the fewest throws of his four games. Minnesota sent five or more at him on 14 of 20 dropbacks, 70 percent, the highest rate of the 30 charted defenses; over the first three weeks opponents sent it at 29. He was catchable on 10 of 17, 59 percent, 20-ninth of 30. On third down he threw five times, completed none, and was sacked once. The costliest snap was second and eight at the Miami 48 in the second quarter, down six: a throw toward Jaylen Wright ruled incomplete, reversed on Minnesota's challenge into a nine-yard sack by Turner, the ball knocked loose and recovered at midfield. −4.7 expected points and 10 points of win probability. Chris Bell drew three throws, all on third down, all deep, none caught. Malik Washington caught all five of his.

The run game was the best thing Miami did, and it got nine carries for the man who made it work. 18 rush plays came to +4.9 expected points, +0.27 a play, against −0.17 a play over the first three weeks. Designed runs gained a hundred 13 yards on 17, 6.6 a run, against 3.2 across the first three games; Minnesota showed six or fewer in the box on nine of them, 53 percent, after 41. Gordon carried nine times for a hundred yards and the touchdown. The scoring drive started at the Miami 45 with a minute 52 left in the third: Washington 17 deep right, Gordon nine, Washington 17 deep left on second and one to the Minnesota 12, and Gordon off left guard for the touchdown with 15 seconds left. 12 to 10. Jaylen Wright carried five times for seven yards.

The turning point came right after it. Minnesota answered with 14 plays and eight and a half minutes, converting third and two, third and six and a fourth and one at the Miami 30, and kicked a 29-yarder for 15 to 10. Miami's next possession was three snaps: Gordon for two, an incompletion for Jalen Tolbert, and on third and eight a deep incompletion for Bell, four points of win probability. Minnesota then drove from its own 19 to the Miami two, and Miami's last timeout went with two oh seven left. Nine possessions: one touchdown, one field goal, the fumble, and four three-and-outs.

The scheme around Willis moved away from its own habits. Motion on 57 percent of charted plays, seventh of the 30, but down from 65 over the first three weeks. Play-action on six of 20 dropbacks, 30 percent, eighth, and up from 25. Willis went four of six for 43 yards off the fake and five of 11 for 42 without it. He was out of the pocket on eight of 20 dropbacks, and neither sack was charted as his own fault, after four of those across the first three games. His contested rate fell to 12 percent, from 26.

The stat lines, scored half-PPR. Gordon, 17 even. Washington, nine and a half. Willis, two and a half, with two runs for 11. For Minnesota, Hockenson scored 18.4 and Jones 14.8.

For anyone joining: Miami lost 27 to 13 in Las Vegas, 35 to 13 in San Francisco, 24 to 10 at home to Kansas City, and 15 to 10 on Sunday, and is oh and four. This is the first week the yardstick is the Dolphins' own 2026. Last week's notebook asked two things. Whether the fifth rusher came back: partway, 23 percent, under the first three weeks' 32, and the sacks came anyway. And whether the run defense would hold a run game under four yards a carry again: no, 4.3 a designed run, though against a Minnesota run game that came in below zero by expected points. What goes in the notebook now: whether a pass defense that allowed +0.48 a play for three weeks and −0.02 on Sunday has turned, or met a passing game that threw short, and whether an offense that ran 37 plays finds a way to give Gordon more than nine carries.

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

_Data via nflverse. Charting: FTN Data via nflverse._