Las Vegas Built a Floor Where There Was None

The 2025 Raiders went 3-14. That part you know. Here's the part that matters for what comes next: our structural win model — which takes a team's point differential and runs it through the fitted points-per-win rate for the season — pegged them at 3.2 wins. The luck delta was -0.2. In other words, Las Vegas was essentially exactly as bad as their record said. There's no hidden competence here, no bounce-back engine buried under a pile of bad fortune. They lost because they were outscored by 191 points, and that number is about as honest a verdict as football produces.

The one-score record adds texture without softening the picture. The Raiders went 2-5 in games decided by one score, which is the kind of record that can go either way in a future year — but it's not a signal when the point differential is already telling the same story. When a team is -191 on the season, they're not being robbed in close games; they're losing the blowouts badly enough that even a few tight wins can't paper over it. This was a franchise at the bottom, not a franchise snakebitten at the top.

The reason to lead with the point differential rather than the record is simple: differential is the most predictive single number in football. Teams that outplay their record tend to rise; teams that underplay it tend to stay put. The 2025 Raiders underplayed theirs by just 0.2 wins. There is no regression gift waiting. The work has to come from the offseason — and that's where the more interesting story begins.

The honest answer to what worked in 2025 is: not much. The EPA profile is negative on both sides of the ball in the ways that hurt most. The offense posted -0.21 EPA per pass attempt and -0.26 EPA per carry — both below zero, both below league average, and the rushing number is particularly ugly given how much of an offense's identity gets tied to the run game when passing isn't working. A team generating negative expected value on nearly every play type is a team that has to get lucky to score, and this one didn't.

The defense was marginally better in the trenches. Opponents produced -0.07 EPA per carry against Las Vegas, which is a genuine positive — it means the Raiders were actually stopping the run at a slightly above-average rate. Pass defense was the problem: opponents generated +0.09 EPA per attempt, which puts the Raiders on the wrong side of league average through the air. The directional story is a front seven that could hold its ground against the ground game while getting carved up by quarterbacks, and that's a coherent picture of a team that needed linebacker and edge help in the worst way.

Situationally, the numbers confirm the narrative. A 40.1% conversion rate on third and fourth down is below the level you need to sustain drives, and a red-zone EPA per play of -0.11 means that even when Las Vegas got into scoring position, they were more likely than not to leave points on the field. Then-coach Pete Carroll — who is no longer with the team — posted a decision index that agreed with the model 82.4% of the time, which is respectable. But his go-for-it rate on fourth down was 17.6% against a model recommendation of 29.4%, and that gap cost Las Vegas roughly 0.235 expected wins in 2025 alone. Over a full Carroll tenure the average was 0.39 expected wins lost per season to decision-making. These aren't catastrophic numbers, but at 3-14, every fraction of a win matters.

The arrivals list is long and the departures list is empty, which tells you something about where this franchise thinks it is. The most significant addition is Tyler Linderbaum, who comes over from Baltimore on a three-year deal at $27M per year — the top of the center market at the time of signing. The market-wins model prices that at 2.24 wins annually, which makes him the most expensive piece of the offseason by a distance. The contract carries a $12M dead-cap figure at exit in 2028, meaning Las Vegas has a real off-ramp if the deal ages poorly, but the entry price deserves a WATCH verdict: this is a team paying 100th percentile money at the position without a production track record in this system yet. If the center market moves substantially past $27M by 2027, it reads as a bargain in hindsight. If it doesn't, they paid to be first.

At edge, Kwity Paye arrives from Indianapolis on three years at $16M annually — 92nd percentile in pay, a WATCH on the same grounds as Linderbaum, with a $16.8M dead-cap number after 2027 that limits flexibility if the production doesn't follow. The linebacker room gets rebuilt with Quay Walker (three years, $13.5M, 98th percentile pay, $2.4M dead at exit) and Nakobe Dean (three years, $12.01M, 97th percentile pay, $12.3M dead after 2027). All four of these deals carry the same disclaimer: graded on price discipline only, because we don't yet have production data in this system. Every single one of them is paying near the top of its positional market.

The most interesting addition on the production ledger is Kirk Cousins, who comes over from Atlanta at $11.3M per year — 0.88 market wins on price — but carries a production-wins figure of 1.8, the only arrival with a meaningful positive production number in the data. That gap between what Las Vegas is paying and what Cousins produced elsewhere is the most straightforward value play in the offseason. Jalen Nailor rounds out the receiving room from Minnesota at $11.68M per year, where his production value is essentially zero (-0.01 production wins), making him the one arrival who doesn't obviously justify his market price. There is no draft class in the brief to chart, and no comp picks projected for 2027.

The net offseason production figure is +1.79 wins on a production basis, driven almost entirely by Cousins. On a market basis — what Las Vegas paid relative to league-average production for those dollars — the number is +7.33, which reflects how aggressively they spent. The honest framing is that the production basis is the more credible number until these players prove themselves in Las Vegas, because market wins assume you're getting what the position is worth at that price. When four of your six major acquisitions have no production data in this system, the 1.79 is the floor estimate and the 7.33 is the ceiling if everything breaks right.

The cap posture has two sides. The positive side: quarterback spending is at just 1.6% of the cap, which is historically the lowest-risk way to carry a roster if the quarterback can play. Brock Bowers sits at the 98th production percentile while consuming only 1.6% of the cap — that's the kind of windfall that lets a team spend elsewhere, and Las Vegas is using that room aggressively. Ashton Jeanty is at 80th percentile production on a 2.6% cap hit, another genuine surplus. The negative side: the edge group is the most expensive in relative terms, at a +1.89 Z-score above league average, anchored by Maxx Crosby's 11.4% cap share. That's a significant concentration of resources at one position, and it has to produce pass-rush value at a level that justifies the spend.

The one clear albatross is Nailor — 71st percentile in pay, 30th percentile in production. He's not a budget-breaker at 2.5% of the cap, but he represents exactly the kind of misallocated spending that compounds quietly. The future proration number of $67.5M means Las Vegas is carrying meaningful dead money from past contracts, which constrains the kind of emergency pivots a rebuilding team sometimes needs. The help-or-hurt verdict is net positive, but conditionally so: if Linderbaum, Paye, Walker, and Dean arrive at their positional market value, this roster is genuinely improved. If two of them don't, the spending looks like addition by activity rather than addition by production.

Here's the arithmetic, stated plainly. The 2025 point differential projects to a structural base of 3.2 wins — a number that reflects how bad last year actually was, not how unlucky. The offseason production additions, led by Cousins and priced conservatively, add 1.79 wins on a production basis. The coaching change from Carroll to Klint Kubliak is effectively a wash in the decision model: Carroll cost 0.39 expected wins per season to decision-making; Kubliak has no head-coaching record, so he's priced at league average, which implies 0.37 expected wins lost — a delta of 0.01 wins. Add it up: 3.2 + 1.79 + 0.01 = 5.0 structural wins for 2026.

That is an accounting projection from the ledger, not a pick, not a prediction, and not a bet. It is the number the math produces when you combine last year's underlying performance with the value added this offseason. The falsifier is straightforward: if Las Vegas wins seven or more games in 2026, this read was too conservative and the market-basis additions delivered closer to their ceiling. If they win three or fewer, the production additions — particularly Cousins — didn't transfer, and the model was too generous. Five wins is the honest middle ground for a team that was genuinely bad, spent genuinely aggressively, and hired a head coach whose value is still an open question.

The brief shows no vacated opportunity on the departure side — Las Vegas didn't lose notable target volume — so the DFS and fantasy case here is built around what the arrivals bring rather than what opened up. The most straightforward name is Bowers. At the 98th production percentile on a 1.6% cap hit, the market already knows he's good; the question is whether a functional quarterback changes his ceiling. Cousins completed a high enough volume of passes in Atlanta to suggest he can feed a tight end who's already elite at creating separation. That's a usage and efficiency argument, not a projection, and it's the kind of thing the price tends to reflect early in the season before the production confirms it.

Nailor is the harder case. He pulled in 53 targets for 100.7 PPR points in 2025 — a per-target rate of 1.9, which is functional but not exceptional. He's paid at the 71st percentile of his position while producing at the 30th. That combination is the definition of a player the fantasy market tends to price on contract value rather than efficiency, which means the real question is whether Cousins elevates the volume enough to make the efficiency irrelevant. On the numbers available, the honest answer is: not obviously. The sleeper case requires a step forward in both target share and per-target value that the current data doesn't support.

The genuine efficiency story on this roster is Bowers, and the aging curves in the brief flag nothing of concern in the skill positions to fade. The one thing worth watching from a DFS standpoint: if Kubliak's offense runs through the run game and Jeanty's 80th-percentile production hold, the backfield becomes the second-most interesting spending decision after Bowers — a player producing at that level on a 2.6% cap hit is exactly the kind of mispricing that surfaces in salary-based formats before the broader market corrects it.

Everything above traces to a live instrument, and the receipts stay public: the GM boards carry Klint Kubliak's decision card, the team's cap posture ($67.5M already charged to future caps), and every contract verdict with its falsifier and review date — including Tyler Linderbaum (WATCH), Kwity Paye (WATCH), Quay Walker (WATCH), Nakobe Dean (WATCH). When the season grades these reads, the grades post whether they flatter us or not.