Lions earned 19-9 in Miami, with a defense doing the work and a thin offense

The verdict

Played like that, the Lions win 92% of the time. The snaps put Detroit ahead of Miami by 7.5 points. The scoreboard read 19-9, a 10-point win, so the result ran 2.5 points past what the film supports. That is an earned win with a small tailwind.

The margin came from the defense and from Miami's mistakes. Detroit's offense did not carry the day.

What the score hid

The Lions averaged 0.01 EPA per snap. EPA, expected points added, measures how much each play moves the scoring outlook of a drive. A figure that close to zero means the offense broke even. Miami's offense averaged minus 0.21 per snap, which is where the gap opened.

Success rate tells the same story. Detroit was successful on 44% of snaps. Miami managed 29%, so the Dolphins lost the down-and-distance battle on more than two of every three plays.

Luck points leaned Detroit's way, 1.6 to minus 1 for Miami. That is a modest tilt, not a rescue.

The biggest swing play came in the first quarter at 5:14. J.Goff's pass intended for A.St. Brown was intercepted by C.Johnson at the Miami 40 and returned 21 yards to the Detroit 39. The play moved win probability by 18 points. It was the loudest snap of the game, and it was a mistake by the Lions that the rest of the afternoon had to absorb.

The calls

Dan Campbell went for it three times and was right by the numbers each time he faced a close call.

In the first quarter, on fourth-and-3 from the Miami 31, the model preferred a field goal. It rated the call a lean, worth 0.3 expected points. Detroit went and converted. In the second quarter, on fourth-and-7 from its own 30, the model preferred a punt, by 0.5 expected points. Detroit went and converted again. In the third quarter, on fourth-and-1 from the 8, the model said go, clearly. The Lions scored a touchdown.

The other four decisions matched the model: field goals at the 10, the 22 and the 14, and a punt on fourth-and-16 from the Detroit 35. Campbell was aggressive where the data says the cost is small and conservative where it says to kick. That is the right pattern.

Miami's flags helped. The Dolphins were penalized four times for swings of 5 to 6 points of win probability: offensive holding in the third quarter at 5:15, offensive holding in the second at 12:27, offensive holding in the first at 5:03, and a false start in the second at 1:11. All four benefited Detroit.

The trend

The Lions are 6-2. The snaps say 4.7 deserved wins, a gap of 1.3 wins, the third-largest luck margin among the 32 teams. Detroit is 4-1 in one-score games and has not been robbed in a single game this year.

The luck is concentrated in one place. Opponent kicking luck accounts for 11.1 of Detroit's 18.6 total luck points, and fumble recovery luck adds 1.4. Those are the margins that tend to even out.

The Chicken Index sits at 20%. Of the five fourth downs where the model recommended going, Campbell went on four. Across 38 graded fourth downs, the Lions have given up 2.3 expected points and have no clear misses. The coach is not the problem.

The real number to watch is net EPA per snap, 0.032, which ranks 13th of 32. That is a good team playing like a middle-of-the-pack one, with a record that looks better than the snaps.

Detroit hosts the New England Patriots in Week 10, Sunday, Nov. 15, at 9:30 AM ET on Fox.