Seahawks at Broncos: A Close Call Under Thursday Night Lights

Denver is the pick at 52.8%, but this is nearly an even game. Our published Thursday winner selections are 2–3 through five weeks, and tonight brings together two teams with much better recent records in primetime.

Seattle and Denver both enter 3–2, with kickoff at 7:15 p.m. CDT on Thursday, October 15, at Empower Field at Mile High. Prime Video carries the game. The Seahawks’ viewing guide confirms the matchup details.

The forecast remains the one issued Tuesday, October 13, at 2:48 p.m. CDT. Thursday brings a fresh market capture and updated football context, not a replacement prediction.

Green identifies the predicted outright winner. The spread and total forecasts below answer separate questions.

Our Thursday record: 2–3

Here is the record using the teams selected in our original published pregame posts, with each result linked to its follow-up.

That is two correct outright picks in five games. Dates follow the U.S. Central calendar, including the Melbourne game. The Seattle–New England season opener was Wednesday, September 9, so it is excluded.

This is a record of what readers saw before kickoff, not a backtest using today’s models. Week 4 showed both the original and rebuilt forecasts; both chose Cleveland. Later retraining does not change these five selections, and this winner record says nothing about spread or total performance.

Dallas’s loss last Thursday is also a useful reminder: all five winner models agreed, but the Cowboys still had only a 59.8% blended probability. Tonight’s 52.8% Denver estimate offers even less separation. Five games are too few to establish a special Thursday strength or weakness.

Two teams comfortable in primetime

Denver’s official preview reports these records entering tonight:

Those are the team’s reported primetime categories and cutoffs. Darnold’s row covers his Seattle tenure, not his entire NFL career. They make tonight an interesting meeting of recent success under the lights; they are not adjustments to the frozen forecast.

The presentation adds some history, too: both clubs will wear throwbacks, recalling their 1977–2001 years together in the AFC West, as Denver’s announcement explains. Seattle’s uniform preview sets up the royal-blue-and-orange matchup.

What the models actually say

Three of the five winner models choose Denver. Their estimates are not unanimous, and none gives the Broncos 60%.

Logistic regression and XGBoost favor Seattle; Elo, Bayesian and Random Forest favor Denver. These are related models, not five independent votes. The blend leaves 47.2% for Seattle and remains raw and uncalibrated, with no additional confidence sharpening.

The separate XGBoost margin and total models project:

The point model therefore also sees a close Denver win. Its total is 2.12 points above 40.5, but a point estimate is not an over probability. Team scores, margin and total are rounded independently from the saved full-precision values.

The separately fitted market forecasts are Denver +3 at 50.8% and Over 40.5 at 53.0%. Neither is a strong signal. The spread blend uses three trained models; the totals blend also includes two neutral 50% proxy contributions. At exactly a three-point Seattle win, the original Denver +3 selection pushes; the export does not supply a separate push probability.

The matchup questions behind the numbers

Can Seattle exploit Denver’s missing coverage pieces without giving possessions away? Denver has ruled out Pat Surtain II and Talanoa Hufanga with ankle injuries, while Riley Moss is questionable with a rib injury. Seattle’s Jaxon Smith-Njigba has no game designation after returning to full participation Wednesday. That makes Denver’s coverage choices a central watch point. Broncos injury report, Seahawks availability report.

Darnold has seven touchdown passes and five interceptions in three games since returning in Week 3. Seattle has discussed using familiar concepts to help him make faster decisions. Denver, meanwhile, produced three takeaways and 13 points off turnovers last week. Sustained drives and short fields can produce the same scoreboard total through very different performances. Seattle’s passing-game report, Denver’s preview.

Can Seattle finish the game with a changed linebacker group? Drake Thomas and Chazz Surratt are out, with Tyrice Knight expected to step in for Thomas. Leonard Williams has no game designation. Zach Charbonnet will not be activated from the physically unable to perform list for this game. All three Denver wins this season required fourth-quarter rallies, putting late possessions and defensive communication on the watch list. Seattle availability, Denver preview.

These are the final injury-report designations rechecked Thursday at 11:56 a.m. CDT. Final game-day inactives have not been verified, so Moss remains a question mark here. The reporting adds context without retroactively changing Tuesday’s numbers.

Thursday’s prices still favor Seattle

The same-book BetMGM comparison now covers three saved observations. All times are CDT; these are snapshots, not live or guaranteed opening and closing prices.

Seattle remains the market favorite. Denver’s moneyline shortened Tuesday to Wednesday, then moved partway back Thursday. Its spread moved from +3 to +2.5 and stayed there; the total stayed 40.5. These observations do not establish why prices changed.

That half-point matters: the saved 50.8% cover probability belongs to Denver +3, not +2.5. A three-point Seattle win pushes the former and loses the latter. The total still matches the original threshold; Thursday’s Over 40.5 price is −108, with Under 40.5 at −110.

Thursday’s report produces no qualifying Seattle–Denver recommendation. Denver +125 has a 44.4% break-even rate, below the model’s raw 52.8%, but that apparent gap does not clear the project’s separate confidence filter. The close winner call and unresolved calibration make it inappropriate to turn the pick into a claim of a validated betting advantage.

What goes on the record tonight

The frozen selections remain Denver to win, Denver +3 and Over 40.5. Friday’s follow-up will grade those original selections, compare the projected 21.64–20.99 with the actual score, and update the Thursday winner record from 2–3. Any push will stay separate.

Tonight opens the fourteen-game slate in the full Week 6 preview. Sunday starts early with Houston–Jacksonville in London and ends with Dallas–Green Bay; the weekly post has the complete picks and their original spread and total thresholds.

Source note: Game 2026_06_SEA_DEN uses the frozen Tuesday forecast and Thursday’s saved market report at prediction-repository revision e2284ab. Historical Thursday results were checked against the original published previews and recaps. Official team reporting was checked Thursday morning. Companion editorial evidence retains full-precision calculations, timestamps, source hashes and the historical pick audit.