Seattle travels to Washington on Sunday as a 7.0-point road favorite, with the total sitting at 40.5. That's a large road number, and the underlying EPA case for it is real.

Table of Contents

1. The two-game EPA snapshot
2. Opponent adjustment (the whole ballgame at this sample)
3. The run game and what it does not tell us
4. What would have to be true for the game to break each way

1. The two-game EPA snapshot

Through two games each (offense = plays run, defense = plays allowed; pass/run splits shown per play type):

On its face this looks like a mismatch: Seattle's defense has been the strongest unit of the four, driven mostly by a strong pass defense number, and Washington's defense — on limited work — has allowed the most per play, mostly through the air. But two-game rates are wobbly, and more importantly, the four teams didn't play the same opponents.

2. Opponent adjustment

Who each side faced matters more than the raw numbers at this sample. Here are the four 2026 opponents' full-2025 EPA/play — the largest reliable sample available — with rank out of 32 teams:

The distortion cuts against Seattle's ledger, not for it. Seattle's two defensive performances came against a top-five 2025 offense (New England) and a bottom-third 2025 offense (Arizona) — a mixed test. Washington's defense faced a middle-of-the-pack Philadelphia offense and the 4th-best 2025 offense in Dallas, which is a meaningful part of why Washington's early defensive number looks so leaky: Dallas was the toughest offensive opponent by a wide margin that either defense has faced through two weeks.

On offense, Seattle drew a top-five 2025 defense (New England) and the league's 26th-ranked defense (Arizona) — a mixed bag but not obviously flattering. Washington drew a top-ten 2025 defense (Philadelphia) and the league's worst 2025 defense (Dallas, 32nd). Washington's offensive number was almost certainly propped up by that Dallas matchup the same way its defensive number was dragged down by it.

Net: the raw two-game picture probably overstates Seattle's defensive edge somewhat and understates how good Washington's defense actually is, because one of Washington's two games came against the team that finished with the NFL's best offensive EPA/play allowed... no — against the offense that finished 4th-best in the league a year ago. That's the main reason a two-game EPA gap shouldn't be read literally in September.

3. The run game

Rushing has told a smaller and noisier story so far:

Seattle's run defense has been the strongest of the four units by this measure, but it's sitting on 47 carries. Washington's ground game has been the more productive of the two through two weeks, which matters because if Seattle's defensive edge does hold up, Washington's most plausible path to keeping the game manageable is staying on schedule on the ground rather than living in third-and-long.

4. What would have to be true

For Seattle to cover the 7: the defensive gap shown in Section 1 has to be mostly real rather than an opponent-quality artifact, and Washington's offensive number — inflated in part by facing the league's worst 2025 defense in week 2 — has to regress toward something closer to average. Seattle's run defense probably needs to keep Washington off schedule; the −0.257 EPA/rush allowed is the least stable of the four defensive numbers here, but it's also the one most directly tied to whether Washington ever gets to a manageable third down.

For Washington to win outright or stay inside 7: the opponent-adjustment story in Section 2 has to be doing real work — Washington's defensive number looks worse than the unit actually is because Dallas was Dallas, and Seattle's offensive number is inflated by an Arizona defense that ranked in the bottom third of the league last year. Washington being at home is worth something on top of that, even if we're not putting a specific point value on it this week.

Where the numbers are least trustworthy: anything that hinges on Washington's defensive EPA (102 plays, one of them against a top-five 2025 offense); anything that treats Seattle's +0.085 offensive EPA as a stable rating rather than a two-game blend against very different opponents; and the run-game splits generally, since 42–63 carries per side is a tiny base to build a game plan on.