← Back to blog

Published on Thu Nov 19 2026 19:35:00 GMT+0000 (Coordinated Universal Time) by cresencio

NFL Week 11 Preview: Ten Unanimous Picks, Three Primetime Tests

Ten of this week’s 13 winner picks have all five models behind them, led by Kansas City at 83.6% and Buffalo at 81.2%. Tonight’s Houston selection is unanimous too, but the separate point model projects a 7.91-point win against a 12.5-point handicap. Agreement on the winner leaves plenty of room for disagreement on the margin.

Week 10 finished 9–5 on winners, 6–8 on saved spreads and 8–5–1 on totals. Random Forest led the winner models at 11–3, and the nine unanimous picks went 6–3. That brings the all-five record since Week 5 to 29–12, or 70.7%. The Week 10 closeout has the full results and probability comparison.

This preview covers November 19–23, including all three primetime matchups. The forecast was issued Tuesday, November 17, at 8:59:11 p.m. CST, using completed games through Week 10. The latest market comparison is DraftKings at 1:26:45 p.m. CST Thursday, with official personnel and weather reporting checked Thursday afternoon. Forecasts remain frozen; Thursday prices and personnel updates provide context around them.

All 13 Week 11 straight-up picks, with Houston, Buffalo, Carolina, Chicago, Dallas, Detroit, Kansas City, Pittsburgh, the Jets, Jacksonville, Denver, San Francisco and Cincinnati highlighted green.

Green marks the predicted outright winner. Opponents are neutral gray; the graphic does not represent spread selections.

All thirteen frozen picks

All kickoffs below are Central Standard Time. “Models” counts Elo, Logistic, XGBoost, Bayesian and Random Forest selecting the ensemble’s side. Spread and total probabilities belong to their displayed original forecast lines, which can differ from the current bookmaker’s prices.

MatchupKickoff, CSTWinner leanModelsOriginal spread leanOriginal total lean
IND at HOUThu Nov 19, 7:15 PMHOU 73.1%5/5IND +12.5 (51.7%)Over 44.5 (52.1%)
MIA at BUFSun Nov 22, noonBUF 81.2%5/5MIA +16.5 (50.02%)Over 45.5 (51.8%)
BAL at CARSun Nov 22, noonCAR 56.0%4/5BAL +2.5 (54.4%)Under 43.5 (51.3%)
NO at CHISun Nov 22, noonCHI 73.4%5/5CHI −11.5 (51.2%)Over 49 (51.1%)
TEN at DALSun Nov 22, noonDAL 63.9%5/5DAL −8 (51.6%)Over 55 (50.3%)
TB at DETSun Nov 22, noonDET 61.4%5/5TB +6.5 (50.6%)Over 46.5 (52.4%)
ARI at KCSun Nov 22, noonKC 83.6%5/5ARI +16 (54.6%)Over 41 (51.0%)
PIT at PHISun Nov 22, noonPIT 59.8%3/5PIT −4.5 (52.6%)Over 38.5 (54.3%)
NYJ at LACSun Nov 22, 3:05 PMNYJ 52.2%3/5NYJ +3 (50.3%)Over 40.5 (51.5%)
JAX at NYGSun Nov 22, 3:25 PMJAX 63.5%5/5JAX −6 (59.8%)Over 46.5 (54.0%)
LV at DENSun Nov 22, 3:25 PMDEN 69.5%5/5LV +10 (50.3%)Under 40 (51.4%)
MIN at SFSun Nov 22, 7:20 PMSF 72.1%5/5SF −3 (56.7%)Over 46 (53.1%)
CIN at WASMon Nov 23, 7:15 PMCIN 59.7%5/5WAS +6 (55.2%)Over 53 (52.1%)

Miami–Buffalo’s spread estimate is effectively even. The tiny Miami lean is 50.02%, with Buffalo at 49.98%; both round to 50.0% at one decimal place. It should not be presented as meaningful cover confidence alongside Buffalo’s much stronger winner estimate.

Schedule check: Pittsburgh–Philadelphia is at noon CST, Jacksonville–Giants at 3:25 p.m., and Sunday night’s Minnesota–San Francisco game is in Mexico City. The table and graphic use the current official schedule; the two afternoon windows differ from the saved fixture file. Eagles schedule, Giants schedule, 49ers game guide.

Ten unanimous winners, different confidence levels

The unanimous group includes every 5/5 winner selection, without a minimum probability cutoff. Its 29–12 record across Weeks 5–10 measures outright winners, not spread results. This week’s ten estimates average 70.1%, but that numerical proximity to the historical 70.7% hit rate does not establish calibration.

Unanimous winnerEnsemble win probabilityThursday DraftKings handicap
Kansas City83.6%KC −15.5
Buffalo81.2%BUF −16.5
Chicago73.4%CHI −11.5
Houston73.1%HOU −12.5
San Francisco72.1%SF −3
Denver69.5%DEN −10
Dallas63.9%DAL −8.5
Jacksonville63.5%JAX −6
Detroit61.4%DET −6
Cincinnati59.7%CIN −6.5

Cincinnati is unanimous at 59.7%, far below Kansas City’s 83.6%. Votes describe agreement; confidence here means the ensemble’s estimated win probability. The models share inputs, so five agreeing models are not five independent confirmations.

The handicap remains its own question. Kansas City is the strongest winner pick, yet the saved spread blend prefers Arizona +16 at 54.6%. Buffalo is second on the winner board, but its cover estimate is essentially a coin flip. Houston, Denver and Cincinnati also have unanimous winner support alongside an opposing spread lean. A favorite winning and an underdog covering are compatible outcomes.

Three primetime previews

Thursday: Indianapolis at Houston

Tonight, November 19, at 7:15 p.m. CST. All five models choose Houston at 73.1%. Our published Thursday winner record is 6–4, following last week’s Giants miss.

The separate point projection is Houston 26.31–Indianapolis 18.40, a 7.91-point margin and 44.71 total. The saved market selections are Indianapolis +12.5 at 51.7% and Over 44.5 at 52.1%. Houston winning by eight would fit the broad margin story while leaving Indianapolis comfortably inside the handicap.

DraftKings holds HOU −12.5 / 44.5 from Tuesday to Thursday. Houston’s moneyline eased −700 → −690, with Indianapolis +500 → +499. Thursday’s spread prices are HOU −12.5 (−106) and IND +12.5 (−114); Over 44.5 is −104, Under −116. The scoring thresholds stayed fixed while the payouts changed.

The main question is how Houston turns its winner advantage into separation. A two-touchdown win clears −12.5, but the point estimate is smaller and neither market probability suggests a confident sweep. At −690, a moneyline wager needs roughly 87.3% wins to break even at that price, above the model’s 73.1% estimate. Selecting Houston as the likely winner is not the same as recommending that price.

Indianapolis has ruled out tackles Bernhard Raimann and Jalen Travis, plus cornerback Charvarius Ward; Sauce Gardner is questionable. Houston is also missing key contributors: Marvin Harrison Jr., Danielle Hunter, Brevin Jordan and Woody Marks are out, while Wyatt Teller is questionable. Those losses put protection, receiving depth and the pass rush into the matchup discussion even though the book’s thresholds stayed fixed. These are final injury-report designations; complete game-day inactive lists remain pending at this afternoon check. Colts final report, Texans final report.

Sunday night: Minnesota versus San Francisco in Mexico City

November 22, at 7:20 p.m. CST, at Estadio Banorte. San Francisco is the designated home team in Mexico City. All five models choose SF at 72.1%, with a separate point projection of 49ers 27.04–Vikings 20.93: SF by 6.10, totaling 47.97. Official game and venue guide.

The saved leans are SF −3 at 56.7% and Over 46 at 53.1%. DraftKings remains SF −3 / 46.5, while moneylines moved SF −162 → −163 and MIN +136 → +138. Thursday’s spread prices are SF −3 (−112) and MIN +3 (−107); Over 46.5 is −109, Under −110.

This is the clearest primetime alignment between the winner selection, projected margin and saved spread lean. A three-point San Francisco win would push −3; four or more would cover. The 53.1% Over estimate still belongs to 46, not Thursday’s 46.5. Mexico City’s venue conditions also matter; a Santa Clara weather forecast would describe the wrong location.

Sunday’s availability is still developing. Wednesday’s report lists Minnesota’s Kyler Murray (hamstring) and Jordan Addison (thumb), plus San Francisco’s Mike Evans (Achilles), as non-participants; T.J. Hockenson (wrist) was limited. Those practice statuses are not final game designations. Minnesota’s updated Wednesday report says Addison will remain on the active roster after thumb surgery, superseding the earlier expectation of injured reserve. That does not establish that he will play Sunday. Official practice report, Vikings’ Addison update.

Monday night: Cincinnati at Washington

November 23, at 7:15 p.m. CST. All five models choose Cincinnati at 59.7%, but the point projection is close: Bengals 27.38–Commanders 25.66, a 1.71-point Cincinnati margin and 53.04 total. The saved selections are Washington +6 at 55.2% and Over 53 at 52.1%.

DraftKings holds CIN −6.5, with the total rising 52.5 → 53.5. Cincinnati’s moneyline moved −298 → −301, Washington +240 → +245. Thursday’s spread prices are CIN −6.5 (−107) and WAS +6.5 (−112); Over 53.5 is −114, Under −105.

The matchup asks whether Cincinnati can turn unanimous winner support into a much larger margin than the point model expects. A six-point Bengals win would push the original Washington +6 and cover a new Washington +6.5. A 53-point finish would go over Tuesday’s 52.5, push the forecast’s original 53 and stay under Thursday’s 53.5. Those are three different settlement outcomes, so the saved percentage cannot simply move with the market.

Monday’s final availability remains pending this Thursday afternoon. The Bengals and Commanders reports will provide the next personnel check before kickoff.

Kansas City’s handicap eases despite the strongest winner estimate. The Chiefs are 83.6% outright against Arizona, with a point projection of 26.76–16.63, a 10.14-point margin and 43.39 total. DraftKings moved KC −16.5 → −15.5, while the original model benchmark was −16. The saved Arizona +16 at 54.6% is a view on the margin, not a prediction that Arizona is the more likely winner.

Buffalo and Chicago also face large margins. Buffalo’s 29.00–17.39 score projection gives it an 11.61-point edge, below the current −16.5, despite an 81.2% winner estimate. Chicago is 73.4% to beat New Orleans, with a separate 27.06–20.47 projection and 6.59-point margin against −11.5. Chicago’s saved cover estimate is only 51.2%. Large winner probabilities should not overshadow those less emphatic margin cases.

Carolina reaches three. Four models choose Carolina at 56.0%; Elo prefers Baltimore. The point model has Carolina 23.13–Baltimore 20.31, a 2.82-point margin. DraftKings moved CAR −2.5 → −3. The saved Baltimore +2.5 at 54.4% would lose on a three-point Panthers win, while a new Ravens +3 would push. The associated price matters as well as the extra half-point.

Carolina’s preliminary Wednesday report lists Bobby Okereke (foot), Darren Waller (knee) and Tommy Tremble (knee) as non-participants, while Bryce Young (foot) practiced fully. Baltimore lists Nnamdi Madubuike (hamstring) as a non-participant. Those are personnel questions to monitor alongside the move to three, not confirmed Sunday absences or proof of its cause. Official Panthers report covering both teams.

Jacksonville pairs unanimity with the strongest saved spread lean. The Jaguars are 63.5% outright and 59.8% at −6, while DraftKings remains −6. The separate point model projects Jacksonville 26.12–Giants 22.93, only a 3.19-point margin. That disagreement is part of the forecast: the winner, cover and margin estimates are separately fitted, and a stronger cover percentage does not make the point model predict a larger win.

Jets–Chargers is the closest winner call. Logistic, XGBoost and Bayesian choose the Jets, while Elo and Random Forest prefer the Chargers. The ensemble lands on NYJ 52.2%, against a market that still favors LAC −3. The saved Jets +3 at 50.3% is only a slight lean; the total moved 40.5 → 41.5.

Pittsburgh gets three votes at Philadelphia. Logistic, XGBoost and Random Forest select Pittsburgh; Elo and Bayesian choose Philadelphia. The ensemble is PIT 59.8%, with a 22.09–18.90 point projection, a 3.19-point margin and 41.00 total. DraftKings remains PIT −4.5, with the total falling 38.5 → 37.5.

Philadelphia confirmed Wednesday that Joe Milton III will make his second start. The Eagles also said DeVonta Smith and Dontayvion Wicks have a chance to return, with practice still determining their availability. That gives the low-total matchup a concrete personnel question: what passing options will Milton have? It does not prove that those updates caused the one-point total move. Official Eagles update.

The full Tuesday-to-Thursday line report

These are DraftKings observations from November 17 at 8:16:16 a.m. → November 19 at 1:26:45 p.m. CST. Tuesday is the first saved project snapshot, not a verified sportsbook opener. Spreads below name the market favorite.

MatchupFavorite spread, Tue → ThuTotal, Tue → Thu
IND at HOUHOU −12.5 → −12.544.5 → 44.5
MIA at BUFBUF −16.5 → −16.545.5 → 44.5
BAL at CARCAR −2.5 → −343.5 → 43.5
NO at CHICHI −11.5 → −11.549.5 → 48.5
TEN at DALDAL −8.5 → −8.554.5 → 55.5
TB at DETDET −6.5 → −646.5 → 45.5
ARI at KCKC −16.5 → −15.541.5 → 40.5
JAX at NYGJAX −6 → −646.5 → 46.5
NYJ at LACLAC −3 → −340.5 → 41.5
LV at DENDEN −10 → −1040.5 → 40.5
PIT at PHIPIT −4.5 → −4.538.5 → 37.5
MIN at SFSF −3 → −346.5 → 46.5
CIN at WASCIN −6.5 → −6.552.5 → 53.5

Three spreads and eight totals changed. Carolina moved to −3, while Detroit eased to −6 and Kansas City to −15.5. Chicago and Dallas illustrate why source labels matter: the saved forecast totals are 49 and 55, respectively, while Tuesday’s DraftKings observations were 49.5 and 54.5. Those differences are not later market moves.

What can drive those changes?

Injury and availability updates can alter scoring expectations. A quarterback change, missing protection or depleted receiving group can affect an offense; a missing edge rusher or defensive back can help the opponent. Expected workload and replacement quality matter, as does whether the news was already anticipated. Philadelphia’s quarterback confirmation and possible receiver returns give readers something specific to follow alongside its lower total.

Weather and venue conditions can change the passing and kicking environment. Wind and rain can influence play selection and ball handling, while a closed roof limits direct outdoor-weather exposure. Forecasts must match the actual venue and game window. Those effects are possible inputs to market decisions, not a reason to apply an arbitrary adjustment to our frozen probabilities.

For tonight’s Houston game, the stadium-area National Weather Service forecast, updated 12:21 p.m. CST Thursday, calls for partly cloudy conditions, northeast wind around 5 mph becoming calm, and patchy fog after 9 p.m. Direct exposure depends on the stadium’s roof position. This supplies actual venue context without attributing the unchanged 44.5 total to weather. NWS stadium-area forecast.

Wagers, a sportsbook’s potential payouts and competitors’ offers also affect prices. Houston’s unchanged −12.5/44.5 thresholds still carry different payouts on the two sides. A move can happen through the number, the attached price, or both. These snapshots show what changed, but contain no betting-volume or bettor-identity evidence establishing a “sharp money” explanation.

The confidence check

The latest completed monitoring receipt is GREEN, with 0.588769 log loss across Weeks 8–10, covering 43 probability-scored games. The wider ensemble winner record since Week 5 is 57–28, or 67.1%, plus one tie; the unanimous subset is 29–12, or 70.7%. Those are different groups, and neither percentage should replace an individual game’s forecast.

The model still uses the raw fallback, T = 1, with calibration unvalidated. Saved market blends retain legacy proxy components; they are not five independently trained spread or total models. Ten agreeing winner selections therefore do not imply ten strong spread bets.

The dashboard’s last betting import uses Tuesday morning’s capture. Thursday’s editorial comparison is not a newly selected portfolio, and settled-bet evidence remains insufficient to calculate ROI. This week’s test is whether the unanimous winner group can keep its recent record moving while the separate margin, spread and total forecasts receive their own accounting.

Source note: the November 19 preview handoff and Thursday market report; the frozen Week 11 forecast and issuance receipt; and the Week 10 recap for historical results. Point scores, margins and totals are rounded independently from full precision. Original lines remain the grading benchmarks.

Written by cresencio

← Back to blog