Houston Built a Defense First, Now Comes the Bill

Twelve wins and a 109-point differential. Run those numbers through the fitted model — structural wins equal 8.5 plus point differential divided by 35.8 — and you get 11.5 structural wins against 12 actual. The luck delta is +0.5, which means almost nothing happened in Houston last year that the points on the scoreboard didn't already explain. That's worth noting because it cuts both ways: the Texans weren't lucky, but they also weren't unlucky. What you saw was largely what they were.

The one-score record was 7-5, which sits on the winning side of a coin flip but doesn't scream hidden fragility or hidden dominance. A team that goes 7-5 in close games while outscoring the rest of the league by 109 points is, in the clearest terms, genuinely good. The structural number — 11.5 — isn't a forecast yet, it's an anchor. Everything that follows in this piece is about whether the offseason moved that anchor up or down.

The defense is the story, and the EPA numbers make it hard to argue otherwise. Opponents produced -0.16 EPA per pass attempt against Houston last year — that's a number that belongs on a short list of the league's genuinely elite pass defenses. The run defense was solid without being spectacular at -0.05 EPA per carry allowed. On a per-play basis, this was a defense that made offenses feel bad about most of their decisions.

The offense is more complicated. A +0.05 EPA per pass attempt is functional — it keeps drives alive and doesn't lose games — but it isn't the kind of number that wins shootouts. The run game was a drag at -0.09 EPA per carry, which matters less when your defense is doing what Houston's was doing but becomes a real problem if the defense ever has a down stretch. The red zone paints the same picture in bolder strokes: -0.14 EPA per play inside the 20 is below average, and that's where the offense left points on the field all season. A 46.3% third-and-fourth-down conversion rate is serviceable but not a weapon.

DeMeco Ryans ran a 9.8% go-for-it rate on fourth down against a model recommendation of 36.1%. His overall decision agreement rate of 67.2% cost Houston an estimated 0.521 expected wins last season. That's not a catastrophic number — it's roughly half a win — but it's real, and it comes entirely from the gap between how often the model says go and how often Ryans actually goes. The defense covers a lot of ground. The decision-making costs a little of it back.

The headline transaction of Houston's offseason isn't an arrival — it's a departure, and it's a significant one. Laremy Tunsil left for Washington at $30.1 million per year (2.51 market wins of value), and Tytus Howard followed him out the door to Cleveland at $22.5 million per year (1.85 market wins). That's 4.36 market wins of offensive line walking out in one offseason. Houston's answer was to spend roughly half that to bring in Braden Smith (RT, from Indianapolis, $10M APY, 0.77 market wins), Wyatt Teller (LG, from Cleveland, $8M APY, 0.59 market wins), and Ed Ingram (RG, from Minnesota, $12.5M APY, 0.98 market wins). Then they used the 26th pick — worth 1.14 chart wins — on Georgia Tech guard Keylan Rutledge. The line got younger and considerably cheaper. Whether it got better is an open question the production wins data can't yet answer for most of the arrivals.

The other major arrival is David Montgomery, who came from Detroit at $8.25 million per year. His production wins number — 0.01 — tells you he's a role player in this model, a reliable presence rather than a difference-maker. Reed Blankenship arrives from Philadelphia at $8.25M APY to reinforce the secondary. Logan Hall comes in from Tampa at $6.875M APY to add interior defensive line depth, and Jadeveon Clowney returns to the league at $5.5M APY. Tim Settle (IDL) left for Washington at $8M APY, the only other meaningful departure on defense.

The draft class totals 4.39 chart wins across eight picks, led by Rutledge (1.14) and Ohio State defensive tackle Kayden McDonald at pick 36 (1.07 chart wins). Michigan tight end Marlin Klein at pick 59 adds 0.80 chart wins. The back half of the class — Oklahoma offensive lineman Febechi Nwaiwu, Clemson linebacker Wade Woodaz, USC safety Kamari Ramsey, Boston College receiver Lewis Bond, and Indiana linebacker Aiden Fisher — contributes a combined 1.38 chart wins at picks where contribution is the reasonable expectation.

Four new contracts carry verdicts. Nico Collins's two-year, $30M-per-year extension rates FAIR PRICE: he's at the 98th percentile of pay against the 89th percentile of production, with an aging multiplier of 0.89 at age 27 and a $21.1M dead-cap exit in 2028 if the relationship sours. The other three are WATCH flags. Will Anderson Jr.'s three-year deal at $50M APY puts him at the absolute top of the edge-rusher market (100th pay percentile) at age 25, with an $18.3M dead-cap exit in 2029 — the falsifier is whether the ED market moves so far past $50M by 2027 that this reads as a discount. Danielle Hunter's one-year deal at $40.1M APY is the 99th percentile of pay for a 33-year-old with no clean exit clause. And Azeez Al-Shaair's three-year, $18M-per-year linebacker deal is at the 99th pay percentile, with $21.7M in dead cap through 2028.

The net offseason number is -0.53 wins on a market basis, 0.01 on a production basis. Those two figures are telling you the same thing in different registers: Houston replaced expensive, established contributors with cheaper, unproven ones and is betting that the difference in price was larger than the difference in quality. That might be right. It's a bet, though, not a certainty.

The offensive line is where that bet is most visible. Tunsil and Howard represented the two largest single-player losses in market-win terms. What replaced them cost less than half as much and carries no production-wins track record in these new roles. The cap posture explains why Houston made this move — the group spending picture shows running backs at -1.49 standard deviations below league average allocation, interior defensive line at -1.16 below average, and linebackers at +1.01 above. Spending on two top-of-market edge rushers and a linebacker while letting the offensive line turn over is a coherent philosophy. It's also the kind of philosophy that gets tested early if C.J. Stroud spends September running for his life.

The edge-rusher contracts are the core tension in this roster. Anderson and Hunter together consume 9.5% of the cap between them (5.2% for Hunter, 4.3% for Anderson), and both are WATCH flags. Hunter's age-33 season at the 99th percentile of pay with no clean exit is the kind of entry that either looks visionary or expensive in November, with little middle ground. Anderson's deal is more defensible — he's 25, the money is locked in, and the aging multiplier question won't answer itself for a few years — but paying 100th-percentile money always requires the player to be right, not just good. The future proration figure of $883.5M reflects how many obligations are already committed forward. Houston is not a team with a lot of financial room to fix mistakes.

The one genuine bright spot in the cap picture is C.J. Stroud at 6.6% of cap — or equivalently, 3.7% on the current-year hit shown in the top-cap table. For a starting quarterback on a team with genuine Super Bowl aspirations, that is a subsidized rate. It won't last forever, but right now it's the structural advantage that makes everything else possible. A team spending top-of-market at edge rusher while paying rookie-scale at quarterback can make that math work. When Stroud's second contract arrives, some of these other commitments will need to be gone.

The arithmetic is straightforward. The structural base from 2025's point differential is 11.5 wins. The offseason adds 0.01 production wins on the ledger. The coaching staff is unchanged, so the decision delta is zero. That produces a 2026 structural projection of 11.6 wins — an accounting output from the ledger, not a betting model and not a pick.

State it plainly: this read is wrong if Houston finishes at or below nine wins. A team that outscored opponents by 109 points in 2025, made modest net changes to the roster, and kept its coaching staff intact has no obvious mechanism for a three-win regression. The falsifier isn't a number pulled from the air — it's the result that the model and the roster composition make genuinely hard to explain. If Houston wins nine games, the offensive line transition almost certainly went badly, at least one of the edge-rusher contracts became a performance story rather than a price story, or Stroud's development stalled. Any of those is possible. None of them is the base case the numbers point toward.

The brief shows no vacated target opportunity from departures — the departures on the ledger were offensive linemen and a defensive tackle, not skill-position players. The DFS and fantasy picture here is more about evaluating what arrived than absorbing a windfall.

David Montgomery is the one arrival with a production profile in the data: 29 targets in 2025, 166.9 PPR points, and 5.76 PPR points per target. Those are the numbers of a legitimate pass-catching back — his per-target efficiency is real, and a team that ran a -0.09 EPA-per-carry offense last year has reasons to lean on the passing game to support its ground attack. If Houston's new offensive line creates any kind of run game improvement, Montgomery's floor rises. His contract at $8.25M APY suggests the market viewed him as a complementary piece. The efficiency says he's worth more than that in the right role.

Collins is the fantasy anchor at receiver, and his FAIR PRICE verdict tells you the market has already caught up to his efficiency. The sleeper case, if there is one on this roster, belongs to whoever emerges as the second receiving option behind Collins in an offense that has repeatedly left points in the red zone. The brief doesn't give us enough on that depth chart to make a confident argument — and stating that gap honestly matters more than filling it with guesswork.

Everything above traces to a live instrument, and the receipts stay public: the GM boards carry DeMeco Ryans's decision card, the team's cap posture ($883.5M already charged to future caps), and every contract verdict with its falsifier and review date — including Will Anderson Jr. (WATCH), Danielle Hunter (WATCH), Nico Collins (FAIR PRICE), Azeez Al-Shaair (WATCH). When the season grades these reads, the grades post whether they flatter us or not.