Detroit Ran the League and Still Went 9-8

Start here, because this number matters more than the record: Detroit outscored its opponents by 68 points in 2025 and finished 9-8. Our fitted points-per-win model — 8.5 base wins plus point differential divided by 35.8 — puts their structural win total at 10.4. The luck delta is -1.4. They were, by the most reliable single measure in football, a ten-win team that won nine.

The one-score record seals it. Detroit went 3-5 in games decided by one score. That's not a scheme problem or a roster problem — late, close games resolve close to randomly for almost every team over a season. A 3-5 mark in those eight games is a coin that landed tails five times. It happens, and then it corrects.

The single most predictive offseason fact for any team is the gap between how it outscored the league and how it outscored its wins column. For the Lions, that gap is 1.4 wins, almost entirely explained by the one-score record. The 2025 season wasn't a warning sign. It was a ledger entry that carries forward into 2026.

Detroit's passing offense was the engine. At +0.18 EPA per attempt, it ranked among the better passing attacks in the league — a number that reflects not just volume but efficiency on a per-play basis. The rushing game was slightly negative at -0.06 EPA per carry, which sounds like a problem until you realize most teams run the ball worse than they pass it, and the Lions' overall offensive structure was built to exploit that gap rather than fight it.

The defense was functional rather than dominant. Opponents generated +0.01 EPA per passing attempt against Detroit — essentially neutral, neither a liability nor an advantage. Against the run, the Lions were slightly better, holding opponents to -0.03 EPA per carry. A defense that keeps the pass game near zero and wins the run game is enough to support an offense operating at +0.18 per attempt. The Lions weren't a defensive powerhouse; they didn't need to be.

The situational numbers fill in the picture. A 43.5% conversion rate on third and fourth down is meaningful — that's sustaining drives rather than punting problems downfield. Red zone EPA of +0.11 per play means they were finishing possessions, not just reaching them. On the coaching side, Dan Campbell's decision index shows he agreed with the model 85.9% of the time on go/kick choices, losing roughly 0.105 expected wins to suboptimal decisions across the season. His actual go-for-it rate of 26.3% trailed the model-recommended 35.4%, which is the quiet story: Campbell is aggressive by reputation, but the model says he still has room left to be more aggressive, not less.

The departures are heavier than the arrivals on paper. Detroit lost five players with meaningful market-win values: linebacker Alex Anzalone (0.63 market wins, to Tampa Bay), running back David Montgomery (0.61 market wins, to Houston), cornerback Amik Robertson (0.55, to Washington), and interior linemen D.J. Reader (0.44, to the Giants) and Roy Lopez (0.35, to Arizona). That's 2.58 market wins walking out the door. Against that, center Cade Mays (from Carolina, 0.62 market wins) and tackle Larry Borom (from Miami, 0.33) arrived for a combined 0.95. The market-basis net is -1.63 wins. On the production basis — what those players actually generated versus what the arrivals are projected to generate — the net is essentially flat at -0.01, because Montgomery's production value came in at just 0.01 wins despite a strong market price.

The draft class is the offseason's most interesting entry. Detroit used the 17th pick on Blake Miller, an offensive tackle from Clemson, worth 1.34 chart wins — a significant investment that signals they're building around the line rather than patching it. Second-round pick Derrick Moore, a defensive end from Michigan, adds 0.87 chart wins. The rest of the class — linebacker Jimmy Rolder (Michigan, round four, 0.36), corner Keith Abney II (Arizona St., round five, 0.27), receiver Kendrick Law (Kentucky, round five, 0.24), and defensive tackles Skyler Gill-Howard (Texas Tech, round six, 0.17) and Tyre West (Tennessee, round seven, 0.15) — adds depth without moving the structural needle much individually. Three comp picks coming in 2027 (two sixth-rounders, one seventh) add another 0.45 chart wins to the future ledger.

Two contract verdicts deserve attention. Jahmyr Gibbs signed a three-year deal at $22.5M per year — 100th percentile in pay, 98th percentile in production, with an aging multiplier of 1.02 at age 25. The model calls it a fair price at current production levels, though there's no clean exit vehicle in the structure. Jack Campbell's four-year extension at $20.25M annually sits at the 99th percentile in pay with production data still accumulating — the model flags it as a watch. There is an exit in 2028 at just $5.2M dead, which is the only thing keeping this from a harder verdict. The price discipline concern is real; the exit clause is the hedge.

Hurt, on the market ledger. Helped, on the production ledger. That's a distinction worth sitting with. The players Detroit lost commanded real contracts elsewhere — but Montgomery, the most prominent departure, produced almost nothing above replacement value by our model despite his $8.25M market price. The gap between what the league paid for these players and what they actually generated suggests Detroit's losses are more cosmetic than structural.

The cap picture reinforces that read. Jared Goff sits at 11.7% of the cap — reasonable for a starting quarterback in a competitive window. Amon-Ra St. Brown at 10.3% is the number that requires the offense to keep performing, because it's a wide receiver eating as much cap as most teams' second-best player at any position. The EDGE group sits at -0.97 standard deviations below the league average in cap allocation, which is the clearest structural vulnerability on this roster: Detroit is spending below market at the position most associated with disrupting the passing game. Miller's arrival at offensive tackle reinforces the line, but a pass rush that can't stress opposing quarterbacks is the logical place for an otherwise well-constructed team to come undone.

There are no albatrosses on this roster. There are two genuine bargains: Gibbs at 1.8% of the cap with 98th-percentile production (before the new deal kicks in fully), and Jameson Williams at 2.6% of the cap with 87th-percentile production. Aging curves raise no red flags — the core is young enough that decline isn't an immediate ledger item. This is a roster in its window, spending intelligently at most positions, with one clear hole at EDGE and one contract (Jack Campbell) worth monitoring.

The arithmetic is straightforward. Detroit's structural base from the 2025 point differential is 10.4 wins. The offseason production net is -0.01 wins — essentially zero. The coaching staff is unchanged, so the decision delta is zero. Structural wins 2026: 10.4.

That is an accounting projection from the ledger, not a betting model and not a pick. It says: if the Lions play football at the same per-play level they played it in 2025, and if the one-score results normalize even modestly from a 3-5 mark, ten wins is where the math points. The falsifier is clear: a finish below nine wins would mean either the passing offense regressed significantly from +0.18 EPA per attempt, the EDGE deficiency became a genuine liability against good quarterbacks in the playoffs or late-season, or the one-score record went negative again — all three of which would be telling you something structural that one season couldn't reveal. A finish at twelve wins, conversely, would mean the close-game correction overshot, which also happens.

Ten wins is the read. Nine-to-eleven is the honest band around it.

David Montgomery's departure is the number that matters here. He carried the ball 158 times and caught 29 targets in 2025, producing 166.9 PPR points. That workload doesn't disappear — it redistributes. The question is who absorbs it, and the brief points directly at one answer.

Jahmyr Gibbs is the sleeper case hiding in plain sight, though the new $22.5M contract will push his fantasy price up in redraft formats. The efficiency argument, however, remains intact: 98th-percentile production at what was still a modest cap share entering the season. With Montgomery's 158 carries and 29 targets now available inside the same offense, Gibbs doesn't need to improve — he just needs the opportunity to scale. That's what a sleeper actually is: a player whose market price hasn't yet caught up with the volume he's about to see.

Jameson Williams is the other name worth flagging. At 87th-percentile production on 2.6% of the cap — and inside a passing offense running at +0.18 EPA per attempt — the efficiency is already there. No aging flags in the data. No departures in the receiver room that dilute his share. The case is quieter than Gibbs's, but the underlying numbers say the market is still leaving something on the table.

Everything above traces to a live instrument, and the receipts stay public: the GM boards carry Dan Campbell's decision card, the team's cap posture ($300.4M already charged to future caps), and every contract verdict with its falsifier and review date — including Jahmyr Gibbs (FAIR PRICE), Jack Campbell (WATCH). When the season grades these reads, the grades post whether they flatter us or not.