Upcoming
DALatGB
MODERATEMoneyline · position
Fair Prob · GB
55.3%
Market
52¢
Edge
+3.8pp
Market line GB by 1.5 · model lean GB 3.2 pts. Fair value is the market line; the lean is graded on its own, never as a position.
2026 Gridiron Edge is live — ratings, projections, and Week 6 model lines vs. Kalshi
Power ratings, MVP & championship edges, and every Week 6 game priced against the live prediction market.
Gridiron EdgeGame Edges
NFL Game Edges — Week 6
Model as of · prices live
Free, always · 2026 season · DAEPA model vs. Kalshi market price · Refreshed daily 8 AM ET · prices every 5 min
This week on the boardNFL Week 6 predictions →NFL Week 7 predictions — early lines →Week 5 results →
Week 6 gamesSeahawks at BroncosTexans at JaguarsBears at FalconsRavens at BrownsTitans at ColtsJets at PatriotsSaints at GiantsPanthers at EaglesSteelers at BuccaneersCardinals at RamsChargers at ChiefsBills at RaidersCowboys at PackersCommanders at 49ers
Where the Model Disagrees
Week 6 · 2026 · the biggest gaps between our model and the market, ranked · free
Edge = our win probability for the backed side (fair value (books) once a game is priced) minus the Kalshi-implied price, in percentage points. Biggest gaps first; only disagreements past our floor for a position are shown — 2 points net of the book's spread. When the two converge, that game simply isn't listed here.
How We Predict NFL Games
Every 2026 team gets a single number on the points-per-game scale — our 2026 power rating — built from 27 seasons of opponent-adjusted EPA, a Bayesian preseason prior that shrinks last season's results toward the mean, blended each week with this season's play-by-play, and per-component offense / defense / special-teams shrinkage. To predict a Week 6 game we take the home team's rating, subtract the away team's, add a standard +2.5 home-field bump, and convert that spread to a win probability through the margin distribution the model fits to actual game results, rather than assuming a bell curve. A Normal fitted to the same data has σ=14.4 — the spread of results around our projected margin, wider than the ≈13.3 you get around a closing line, because a closing line is a better centre than our rating difference is. The empirical version differs most in the middle, where NFL margins pile up on 3 and 7 and a bell curve does not.
The edge is that number minus the market's. A prediction-market contract settles at $1 or $0, so its price in cents is an implied probability, which makes the comparison direct. We publish the disagreement, not a recommendation.
MODERATE (2)
DALatGBUpcoming
WASatSF
MODERATEMoneyline · position
Fair Prob · WAS
25.3%
Market
22¢
Edge
+3.8pp
Market line SF by 7.5 · model lean WAS 0.5 pts. Fair value is the market line; the lean is graded on its own, never as a position.
2 Week 6 Games Still to Play
Sun, Oct 18, 8:20 PM EDT
DALatGBFair value
55.3%
GB win
Kalshi
50¢
GB YES
Edge
+5.8pp
fair − mkt
BUY YES GB @ 50¢
Power ratings: DAL -0.80 · GB 1.38. Projected margin +4.7 for GB (HFA +2.5). Model line GB by 4.7 · market GB by 1.5 · lean GB 3.2 pts (WATCH). Win probability is priced off the market-anchored line; a position needs 2pp net of the spread and is withheld past 7pp. Kalshi: DAL 51¢ · GB 50¢.
Mon, Oct 19, 8:15 PM EDT
WASatSFFair value
74.7%
SF win
Kalshi
79¢
SF YES
Edge
-3.8pp
fair − mkt
BUY YES WAS @ 22¢
Power ratings: WAS -1.80 · SF 2.81. Projected margin +7.1 for SF (HFA +2.5). Model line SF by 7.1 · market SF by 7.5 · lean WAS 0.5 pts. Win probability is priced off the market-anchored line; a position needs 2pp net of the spread and is withheld past 7pp. Kalshi: WAS 22¢ · SF 79¢.
Every 2026 NFL Week 6 Matchup: Our Model vs. the Kalshi Market
Seattle Seahawks at Denver Broncos — Week 6 Prediction
The Denver Broncos host the Seattle Seahawks. Our 2026 power rating has DEN at +1.30 and SEA at +3.92 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 0.1-point projected margin in favor of Seattle Seahawks. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 41.3% on DEN and 58.8% on SEA. This game has been played; its Kalshi market has settled. Without a live orderbook we can't compute an edge — check back closer to kickoff.
Houston Texans at Jacksonville Jaguars — Week 6 Prediction
The Houston Texans and Jacksonville Jaguars meet on a neutral field. Our 2026 power rating has JAX at +3.17 and HOU at +0.77 on the points-per-game scale. With no home-field adjustment on a neutral site, that's a 2.4-point projected margin in favor of Jacksonville Jaguars. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 58.8% on JAX and 41.3% on HOU. This game has been played; its Kalshi market has settled. Without a live orderbook we can't compute an edge — check back closer to kickoff.
Chicago Bears at Atlanta Falcons — Week 6 Prediction
The Atlanta Falcons host the Chicago Bears. Our 2026 power rating has ATL at +0.02 and CHI at +2.17 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 0.4-point projected margin in favor of Atlanta Falcons. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 44.7% on ATL and 55.3% on CHI. This game has been played; its Kalshi market has settled. That's a 44.2pp gap — past our 7pp ceiling. A disagreement that large with a liquid price is more likely our error than the market's, so no call. The model's own line is ATL by 0.4 against a market line of CHI by 1.5 — a 1.9-point lean toward ATL, graded on its own, not as a position.
Baltimore Ravens at Cleveland Browns — Week 6 Prediction
The Cleveland Browns host the Baltimore Ravens. Our 2026 power rating has CLE at -0.89 and BAL at +1.76 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 4.4-point projected margin in favor of Cleveland Browns. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 39.5% on CLE and 60.5% on BAL. This game has been played; its Kalshi market has settled. Without a live orderbook we can't compute an edge — check back closer to kickoff.
Tennessee Titans at Indianapolis Colts — Week 6 Prediction
The Indianapolis Colts host the Tennessee Titans. Our 2026 power rating has IND at -0.72 and TEN at -4.92 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 6.7-point projected margin in favor of Indianapolis Colts. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 71.7% on IND and 28.3% on TEN. This game has been played; its Kalshi market has settled. Without a live orderbook we can't compute an edge — check back closer to kickoff.
New York Jets at New England Patriots — Week 6 Prediction
The New England Patriots host the New York Jets. Our 2026 power rating has NE at +3.87 and NYJ at -2.31 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 8.7-point projected margin in favor of New England Patriots. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 74.7% on NE and 25.3% on NYJ. This game has been played; its Kalshi market has settled. Without a live orderbook we can't compute an edge — check back closer to kickoff.
New Orleans Saints at New York Giants — Week 6 Prediction
The New York Giants host the New Orleans Saints. Our 2026 power rating has NYG at -0.98 and NO at -3.04 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 0.3-point projected margin in favor of New York Giants. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 48.2% on NYG and 51.8% on NO. This game has been played; its Kalshi market has settled. That's a 43.7pp gap — past our 7pp ceiling. A disagreement that large with a liquid price is more likely our error than the market's, so no call. The model's own line is NYG by 0.3 against a market line of NO by 0.5 — a 0.8-point lean toward NYG, graded on its own, not as a position.
Carolina Panthers at Philadelphia Eagles — Week 6 Prediction
The Philadelphia Eagles host the Carolina Panthers. Our 2026 power rating has PHI at -0.07 and CAR at -0.55 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 1.3-point projected margin in favor of Carolina Panthers. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 39.5% on PHI and 60.5% on CAR. This game has been played; its Kalshi market has settled. That's a 39.0pp gap — past our 7pp ceiling. A disagreement that large with a liquid price is more likely our error than the market's, so no call. The model's own line is CAR by 1.3 against a market line of CAR by 3.0 — a 1.7-point lean toward PHI, graded on its own, not as a position.
Pittsburgh Steelers at Tampa Bay Buccaneers — Week 6 Prediction
The Tampa Bay Buccaneers host the Pittsburgh Steelers. Our 2026 power rating has TB at -0.57 and PIT at -0.54 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 1.8-point projected margin in favor of Pittsburgh Steelers. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 41.3% on TB and 58.8% on PIT. This game has been played; its Kalshi market has settled. That's a 3.3pp gap — past our 7pp ceiling. A disagreement that large with a liquid price is more likely our error than the market's, so no call. The model's own line is PIT by 1.8 against a market line of PIT by 2.5 — a 0.7-point lean toward TB, graded on its own, not as a position.
Arizona Cardinals at Los Angeles Rams — Week 6 Prediction
The Los Angeles Rams host the Arizona Cardinals. Our 2026 power rating has LAR at +2.59 and ARI at -4.40 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 9.5-point projected margin in favor of Los Angeles Rams. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 86.6% on LAR and 13.4% on ARI. This game has been played; its Kalshi market has settled. That's a +2.1pp gap in favor of LAR — above our 2pp floor (net of the book's spread). Buy LAR YES at 85¢. The model's own line is LAR by 9.5 against a market line of LAR by 12.5 — a 3.0-point lean toward ARI, graded on its own, not as a position.
Los Angeles Chargers at Kansas City Chiefs — Week 6 Prediction
The Kansas City Chiefs host the Los Angeles Chargers. Our 2026 power rating has KC at +0.36 and LAC at -2.58 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 5.4-point projected margin in favor of Kansas City Chiefs. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 74.7% on KC and 25.3% on LAC. This game has been played; its Kalshi market has settled. That's a -3.8pp gap pricing KC too high — above our 2pp floor (net of the book's spread). Buy LAC YES at 22¢. The model's own line is KC by 5.4 against a market line of KC by 7.5 — a 2.1-point lean toward LAC, graded on its own, not as a position.
Buffalo Bills at Las Vegas Raiders — Week 6 Prediction
The Las Vegas Raiders host the Buffalo Bills. Our 2026 power rating has LV at -3.21 and BUF at +2.42 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 3.1-point projected margin in favor of Buffalo Bills. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 32.9% on LV and 67.1% on BUF. This game has been played; its Kalshi market has settled. Gap is inside our 2pp floor (net of the book's spread) — no position. The model's own line is BUF by 3.1 against a market line of BUF by 5.0 — a 1.9-point lean toward LV, graded on its own, not as a position.
Dallas Cowboys at Green Bay Packers — Week 6 Prediction
The Green Bay Packers host the Dallas Cowboys. Our 2026 power rating has GB at +1.38 and DAL at -0.80 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 4.7-point projected margin in favor of Green Bay Packers. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 55.3% on GB and 44.7% on DAL. Kalshi's prediction market is pricing GB YES around 50¢ and DAL YES around 51¢. That's a +5.8pp gap in favor of GB — above our 2pp floor (net of the book's spread). Buy GB YES at 50¢. The model's own line is GB by 4.7 against a market line of GB by 1.5 — a 3.2-point lean toward GB, graded on its own, not as a position.
Washington Commanders at San Francisco 49ers — Week 6 Prediction
The San Francisco 49ers host the Washington Commanders. Our 2026 power rating has SF at +2.81 and WAS at -1.80 on the points-per-game scale. With the game model's +2.5 home-field adjustment, that's a 7.1-point projected margin in favor of San Francisco 49ers. Anchored to the market line and converted through the market's own win curve (σ=11.3), the model lands at 74.7% on SF and 25.3% on WAS. Kalshi's prediction market is pricing SF YES around 79¢ and WAS YES around 22¢. That's a -3.8pp gap pricing SF too high — above our 2pp floor (net of the book's spread). Buy WAS YES at 22¢. The model's own line is SF by 7.1 against a market line of SF by 7.5 — a 0.5-point lean toward WAS, graded on its own, not as a position.
Moneyline positions are priced off a market-anchored fair value (the market line) through the market's own win curve. Edge = fair probability minus Kalshi price, in percentage points. A position needs 2pp net of the book's spread: HIGH ≥3.5pp · MODERATE ≥2pp; past 7pp we make no call. The model's own line vs the market line is the lean, in points (HIGH ≥5 · MODERATE ≥3.5 · WATCH ≥2), graded separately. Spread/total β and σ are heuristic pending backtest calibration. Full methodology →
MCP · nfl_edge
Go Pro · 26% off your first year
Every NFL game priced against the live Kalshi line, every week — $150/yr
Pro is the layer around the free game edges: fair value against the live Kalshi price on every market, edge alerts in Discord the moment a gap opens, and the full win-total, championship and projection boards.
Go Pro →
$111 for the first year, then $150/yr. Code applies automatically. Ends Dec 31. Monthly is $14.99/mo.
Every signal graded in public, losers left on the board. See the record →
Same model, inside your AI agent
The engine behind these numbers is callable from Claude, ChatGPT or Grok through our MCP server — every Kalshi prop this week priced venue by venue, spread and win-total ladders against the whole distribution, correlation-aware same-game combos, and the weekly game edges. The free tools answer with no key and no account.
See the in-season NFL tools →NFL game edges vs Kalshi — FAQ
What is a Gridiron Edge, and how is it different from a Kalshi price?
A Kalshi price is the market-implied probability of an outcome — a 58¢ moneyline contract means the market thinks that team wins 58% of the time. Our Gridiron Edge is our own DAEPA power-rating model’s win probability for the same game. The "edge" is the gap between the two, in percentage points: when our model says 66% and Kalshi is pricing 58%, that’s an 8-point edge on the YES side.
How big does an edge have to be before you flag it?
In-season the published line is anchored to the market — the market line — and converted through the win curve the market itself prices, so a position only publishes when that fair value clears the Kalshi price by 2 percentage points net of the book’s bid/ask spread (MODERATE), or 3.5pp (HIGH). Past 7pp we make no call: a gap that large against a liquid price is more likely our error. The model’s own line is still shown against the market line, in points, as a lean — graded on its own record, never as a position. Expect a few positions a week, not a dozen.
How accurate is the model behind these edges?
The Gridiron Edge model backtests to 59.6% straight-up accuracy across 3,266 NFL games from 2013–2024, walk-forward, with no market line among the inputs — the model never sees future results when pricing a game. Accuracy climbs with the model’s own confidence, to 71.5% on the games where it flags an edge above 20%. The market’s closing line is still the stronger raw predictor; the model earns its keep by disagreeing with it in a measurable way. An edge is a probabilistic lean, not a lock.
Can you trade these NFL game markets on Kalshi?
Yes. Kalshi is a CFTC-regulated exchange available in all 50 states. Each game market is a yes/no contract that settles at $1 or $0 — a position on an outcome, not a wager against a sportsbook line. Contracts trade 1¢–99¢ as implied probabilities with no vig.