NFL Week 9 Preview: Strong Favorites, Seattle’s Reversal and Three Primetime Tests

San Francisco at 72.7% and Kansas City at 70.3% lead our Week 9 winner forecasts. Both lay ten points in Thursday’s BetMGM snapshot, and neither has a similarly strong cover estimate. Seattle’s spread has reversed sharply amid quarterback news. Tonight, Jacksonville visits a Baltimore team preparing a first-time starter; Monday’s Buffalo–Minnesota winner forecast is effectively even.

This is one expanded preview for all 15 games, November 5–9, including a closer look at the three primetime matchups. The forecasts were issued Tuesday, November 3, at 10:48:44 a.m. CST, with training through Week 8. They remain frozen. Our latest market comparison is Thursday at 7:42:41 a.m. CST, and the personnel check is from approximately 8:10 a.m. CST. New prices and injury reports provide context; they are not manual revisions to those forecasts.

Week 8 ended 10–4 on winners, 8–6 against the saved spread and 5–9 on totals, including a 2–1 winner record, 3–0 spread sweep and 2–1 total record in primetime. The full recap explains those results. This week starts with the same distinction: choosing the winner, covering a handicap and clearing a scoring total are separate questions.

Green marks the predicted outright winner; opponents are neutral gray. Minnesota’s 50.02% is a rounding-visible lean in an essentially even game.

All fifteen frozen picks

All kickoff times are Central Standard Time. “Models” counts the five base winner models selecting the aggregate’s side; they are weighted differently and share information, so this is not an independent majority vote. LAR means the Rams; LAC means the Chargers. Spread and total percentages apply only to the displayed original lines. These are forecast leans across the slate, not a refreshed list of qualifying bets.

Minnesota’s unrounded estimate is 50.0188%, against Buffalo’s 49.9812%. That is effectively a coin flip. Denver and Cleveland also deserve restraint: each has only two base models on its side, but their weighted probabilities narrowly tip the aggregate toward those teams.

Three primetime games, three different tests

Thursday: Jacksonville at Baltimore

Tonight, November 5, at 7:15 p.m. CST. Jacksonville arrives after beating Indianapolis 42–7; Baltimore lost 36–10 at Buffalo. Four winner models favor Jacksonville, with Elo choosing Baltimore. The aggregate gives the Jaguars 56.4%, a moderate lean alongside a much stronger market favorite.

The separate point model projects Jacksonville 24.03–Baltimore 22.78, a 1.25-point Jacksonville margin and 46.82 total. Its original market leans are Baltimore +9 at 53.7% and Over 42.5 at 54.2%. The projected margin falls well short of both Tuesday’s nine-point handicap and Thursday’s ten. These outputs come from different models, rather than one exact-score prediction that generates every percentage.

BetMGM’s sequence is JAX −9 → −10 → −10, with the total falling 42.5 → 40 → 39.5. Jacksonville’s moneyline becomes more expensive at −475 → −525 → −550; Baltimore moves +360 → +400 → +425. Thursday’s spread prices are JAX −10 (−108) and BAL +10 (−110). The saved cover and total probabilities have not been recalculated for +10 or 39.5.

Baltimore’s quarterback situation is the central football question. Lamar Jackson (right thumb) and Tyler Huntley (toe) are ruled out, and the Ravens announced Joe Fagnano will make his first NFL start. He has no game designation after full participation Tuesday and Wednesday. Baltimore also ruled out safety Jaylinn Hawkins; Nate Wiggins and Adam Randall are questionable. Those are final injury designations, while tonight’s complete inactive lists remain pending. Ravens’ starting-quarterback announcement, official injury report.

Jacksonville has its own losses: DaVon Hamilton and Eric Murray are out for the season, and Travon Walker and Bhayshul Tuten are questionable. My matchup question is whether Baltimore can sustain drives and support its inexperienced starter against a defense missing important pieces, rather than assuming last week’s 35-point Jacksonville win repeats on the road. Those absences pull the analysis in different directions; they do not justify assigning the entire total decline to Baltimore’s quarterbacks. Jaguars’ injury update, both teams’ final designations.

Our published Thursday winner record is 5–3 after Green Bay’s Week 8 win. Tonight tests a moderate model lean against a double-digit market handicap, with a personnel situation the frozen forecast does not update as news arrives.

Sunday night: Tampa Bay at Chicago

November 8, at 7:20 p.m. CST. All five winner models choose Chicago, with an aggregate 60.5% estimate. The Bears just won 34–17 at Seattle; Tampa Bay shut out Atlanta 16–0. Both results are already included in the training inputs.

The point projection is Chicago 24.97–Tampa Bay 21.86, a 3.10-point margin and 46.83 total. That provides a useful contrast with Chicago’s unchanged −6.5 market handicap. Our separate spread lean is Tampa Bay +6.5 at 54.4%: Chicago winning and Tampa Bay covering are compatible outcomes. The saved total lean is Over 46.5 at 52.4%.

BetMGM’s total moved 46.5 → 47.5 → 47.5, with Chicago’s moneyline −275 → −285 → −285. Thursday offers CHI −6.5 (−105) and TB +6.5 (−112). Even the unchanged total has a price change: Over 47.5 moved from −105 Wednesday to −110 Thursday, while Under moved from −112 to −108. A $100 profit on that over now requires $110 risked instead of $105. The 52.4% estimate still belongs to Over 46.5.

Wednesday’s practice report is preliminary. Chicago’s Kyle Monangai (thumb/toe) and Grady Jarrett (groin), and Tampa Bay’s Emeka Egbuka (foot) and Yaya Diaby (knee), did not practice; Baker Mayfield (right thumb) participated fully. A missed practice is not a ruling-out. Receiver and running-back availability could affect sustained offense, while defensive absences could create scoring opportunities. Final Sunday designations are still ahead. Official Bears report covering both teams.

The game to watch is whether Chicago can turn the win lean into a full-touchdown margin. The models favor the Bears outright, but the point projection and original spread selection both leave room for a closer game.

Monday night: Buffalo at Minnesota

November 9, at 7:15 p.m. CST. Buffalo’s 36–10 win over Baltimore and Minnesota’s 19–15 loss at Detroit lead into the week’s closest winner forecast. Three models narrowly prefer Minnesota; Bayesian and Random Forest prefer Buffalo. The aggregate split is MIN 50.0188% / BUF 49.9812%.

The separate point projection actually gives Buffalo a narrow lead: Buffalo 24.57–Minnesota 23.98, a 0.58-point margin and 48.55 total. Scores, margin and total are rounded independently from full precision. Neither this point estimate nor Minnesota’s tiny winner lean suggests a decisive favorite.

BetMGM holds Buffalo −3.5 across the three captures. Its price improved from −102 Tuesday to +100 Wednesday and remains there Thursday; Minnesota +3.5 is −118. Moneylines remain BUF −190 / MIN +155, and the total is 46 → 45.5 → 45.5. Thursday’s prices are Over −108 and Under −110. There has been no further change in these listed markets since Wednesday.

The original spread selection is Buffalo −3.5 at 55.0%, and the total selection is Over 46 at 54.1%. Buffalo’s cover probability exceeds its roughly 50% win probability, even though covering −3.5 requires winning. The separately fitted winner and spread models therefore do not form a coherent joint probability forecast. We show the disagreement rather than treating the two outputs as mutually supporting evidence of an edge. The total percentage likewise cannot be moved to 45.5 without a new estimate.

Monday’s current injury designations were not yet available in the official reports checked Thursday morning. The viewing question is whether Buffalo creates more than a field goal of separation in a matchup the winner model sees as even; the final availability picture remains open.

Sunday’s stories beyond primetime

San Francisco and Kansas City lead outright, with narrower spread cases. All five models agree on each winner, but the saved spread estimates are SF −10 at 52.3% and NYJ +9 at 51.4%. The point model projects San Francisco 26.43–Las Vegas 23.25 and Kansas City 26.88–Jets 23.61: margins of 3.18 and 3.28 points, respectively. Both market favorites currently lay ten. The task is creating separation, not merely winning.

There are five unanimous winner picks: San Francisco, Kansas City, Detroit, Chicago and Cincinnati. Their probabilities range from Cincinnati’s 56.9% to San Francisco’s 72.7%, so model agreement does not make them equally strong selections. Detroit is 61.9% at Miami, but its −7 cover estimate is only 51.8%; that handicap and the 50.5 total are unchanged across all three captures.

Seattle’s reversal has new quarterback context. BetMGM moved SEA −11 → −12 → −8.5, while the total fell 44.5 → 41.5 → 39.5. Seattle announced Wednesday that Sam Darnold will miss the Arizona game with a quad injury. Drew Lock is expected to start but is managing an elbow injury, with Jalen Milroe also receiving preparation reps. That news is relevant to offensive expectations; the snapshots alone cannot establish how much of either move it explains. Seahawks’ November 4 announcement.

Tuesday’s frozen Seattle 61.7% winner estimate has not been recomputed for that announcement. Arizona +11 at 57.0% is the strongest saved spread lean, but Arizona now receives only 8.5 points, losing 2.5 points of protection relative to that benchmark. Under 44.5 at 51.6% also cannot be carried down to Under 39.5. This is where a dated forecast and newer information most visibly need to sit side by side.

The Rams’ handicap keeps shrinking. Los Angeles is a 59.0% winner lean at Washington, while the market moves LAR −6 → −5.5 → −4.5 and the total 47.5 → 47.5 → 48. The original LAR −6 estimate is 55.6%; a smaller current handicap changes the required margin, not the stored percentage. Both teams won last week: the Rams beat the Chargers 24–9 and Washington beat Philadelphia 27–10.

The three-point boundary matters for the Giants and Patriots. New York moves −3.5 → −3.5 → −3 at Philadelphia, while New England moves −2.5 → −3 → −3.5 against Green Bay. A three-point Giants win would now push instead of lose; a three-point Patriots win would lose Thursday’s spread after covering Tuesday’s. Their winner estimates are NYG 56.5% and NE 53.4%. Dallas remains a 56.9% winner lean at Indianapolis as its handicap eases from −6 to −5.5; its 51.7% cover estimate belongs to the original −6.

Houston–Chargers has the strongest saved total lean: Over 40.5 at 56.7%. Thursday’s line is 41. Houston is 60.0% to win, with LAC +5 at 55.9% the separate spread preference. The point model is much closer, Chargers 21.99–Houston 21.81, illustrating why its score estimate is not interchangeable with the winner forecast.

Sunday starts early with Cincinnati–Atlanta in Madrid at 8:30 a.m. CST. All five models choose Cincinnati, but the 56.9% aggregate is modest and Over 52 at 50.2% is barely a lean. Denver–Carolina (DEN 51.8%) and Cleveland–New Orleans (CLE 50.4%) are still closer winner calls. Cleveland’s small preference sits against market favorite New Orleans −3; that is disagreement to watch, not a confident upset prediction. NFL Week 9 schedule and injury reference.

The full Tuesday-to-Thursday line report

Every comparison uses BetMGM, at Tuesday, November 3, 7:52:42 a.m. → Wednesday, November 4, 11:33:55 a.m. → Thursday, November 5, 7:42:41 a.m. CST. Tuesday is our first saved weekly observation, not a verified sportsbook opener. Thursday’s prices are a morning snapshot, not closing lines.

The spread below is the designated home team’s handicap: a plus sign means that team receives points. Totals are combined points. Cincinnati–Atlanta is a neutral-site game with Atlanta designated home.

Five spreads and five totals changed since Wednesday; nine spreads and thirteen totals differ from Tuesday. Seattle’s 3.5-point handicap reversal is the largest Wednesday-to-Thursday spread move. Its total has fallen five points since Tuesday; Jacksonville–Baltimore is down three. Those changes alter the available bets without changing the forecast record.

What could be moving the lines?

Injuries can change both expected scoring and the balance between teams. Seattle’s quarterback change is a concrete development to put beside its smaller handicap and lower total. Baltimore’s two unavailable quarterbacks raise offensive questions, while Jacksonville’s defensive losses could help the opposing offense. The timing, what bettors already anticipated, and replacement quality all matter. These are plausible influences, not proof that a particular report caused a measured move.

Weather is relevant when it changes passing, ball handling or kicking conditions. The retrievable National Weather Service discussion, issued Wednesday morning, describes a front near the I-95 corridor Thursday night and rain chances extending into Friday. That puts possible wet conditions on tonight’s watch list. It is not a precise kickoff forecast, and it does not establish that weather drove the three-point total decline. Dated NWS Baltimore/Washington discussion.

Bets and sportsbook decisions can move either the number or its price. Books respond to wagers, potential payouts and competitors’ offers. Chicago’s total price changed even though the threshold stayed at 47.5; Seattle’s threshold itself changed substantially. We have no ticket counts, stake totals or bettor identities in these captures, so there is no basis to label a move “sharp money.” A changed quote is evidence of the market’s offer, not proof of its cause or profitability.

The confidence check

The latest monitoring is Tuesday’s GREEN assessment, with 0.68959 log loss over 42 games from Weeks 6–8. The season winner record is 73–47, or 60.8%, excluding one tie and retaining the original Week 4 forecasts. GREEN means the available measures cleared the project’s warning thresholds; it does not validate every displayed percentage or establish a betting return.

The probabilities remain raw and uncalibrated. The totals aggregate also retains constant 0.5 Elo/Bayesian proxy components alongside trained total models. Small leans near 50% and separately fitted point estimates should not become confident exact-score predictions. Buffalo–Minnesota’s incompatible win and cover probabilities are a specific limitation, not something to hide in the table.

The dashboard still reflects Tuesday’s 81 value offers and 11 recommendations. Thursday’s line check did not rerun portfolio selection, so those are not Thursday recommendations at newly observed prices. Settled-bet ROI remains unavailable.

Week 9 begins with San Francisco and Kansas City as the strongest winner picks, Seattle’s frozen forecast facing a changed quarterback situation, and three primetime games with distinct questions about margin and scoring. Every original spread and total above stays attached to its original probability for grading.

Sources: the issued Week 9 forecasts, pregame issuance receipt, Tuesday monitoring record, and the project’s November 5 preview handoff and Thursday line report. Point estimates are rounded independently from full precision. Official personnel reporting is linked where discussed; final Thursday inactives and later-week game designations were pending at the morning check.