Sunday, Sep 20, 1:00 PM ET
Nothing to shop yet
Only DraftKings quotes this game so far, so there is no second number to compare it against. Books further out post one at a time; this fills in as kickoff gets closer.
Both prices with the bookmaker's margin removed. This is the market's number, not ours — we don't forecast NFL games, and we publish why.
Teams the market priced at 50–55% won 341 of 644 games since 1999.That is almost exactly what the price claimed — this market prices accurately.
The two prices add up to 104.3%. Everything over 100% is the book's edge, taken whichever side you pick.
CAR · 16 games, 478 attempts · 2025 figures
No games played yet this season — every figure here is from 2025.
29.9 attempts per game · team runs 57 plays at 53% pass · sacked on 5.3% of dropbacks
6.3 yards per attempt · 63.6% completions · -0.044 EPA per dropback
ATL has allowed 7.2 yards per attempt across 17 games
Following the DraftKings total · 14 prices on record
Opened 43.5 on Wed, Jul 29, 53 days before kickoff.
13 reprices since, 13 points of movement in all. It sits 44.5 — 1 point higher than the open. It ranged from 43.5 to 44.5 along the way, retracing 12 points of that movement.
Half of the movement was over by Sat, Aug 15, 8 AM ET.
1 point at once — 43.5 to 44.5 on Thu, Aug 6, 5 AM ET.
Every sentence above is arithmetic on prices we recorded and the times we recorded them. Nothing here says where this number goes next.
Counts from 6,967 regular-season games since 1999. No model, no forecast — this is the history of the number itself, not a claim about this game.
Favourites of 1.5 covered 99 of 194 games.
The range covers a coin flip.
They won outright in 104 of 194.
Totals of 44.5 went over 129 of 294 times.
The range clears a coin flip.
Home handicap. Dashed rules are the key numbers.
Only one price recorded so far — nothing to plot until it moves.
Over line.
The books haven't posted player props for this game yet — normal this far out. The archive is already watching for them: the moment a price appears it is captured, kept forever, and becomes loggable here.
When they land you'll get the same book you get on the game lines — closing-line value, a letter grade and a good-bet verdict on every prop you log. No projections: our prop model failed its own pre-registered calibration gate, and we don't ship numbers that failed.
Every point above is a price we observed and stored, with the time we saw it. Nothing on this page is a prediction, a rating, or a recommended bet.
Every row is a rate measured over games already played, never a probability — the same components behind each quarterback's Index below. No games played yet this season for at least one side, so those figures are last season’s. Our passing model failed its calibration test; we publish that on the accuracy page.
Expected volume from the inputs above. Unvalidated — it is arithmetic on what has already happened, not a forecast that has been scored against outcomes.
The market has his team at 21.5 points implied.
The index ranks this quarterback and this matchup against every qualifying passer, over games already played. It is not a probability and not a bet — our passing-yards model failed its calibration test, and we publish that on the accuracy page.
ATL · 9 games, 276 attempts · 2025 figures
No games played yet this season — every figure here is from 2025.
30.7 attempts per game · team runs 60 plays at 53% pass · sacked on 4.5% of dropbacks
7.2 yards per attempt · 60.1% completions · 0.018 EPA per dropback
CAR has allowed 7.2 yards per attempt across 17 games
Expected volume from the inputs above. Unvalidated — it is arithmetic on what has already happened, not a forecast that has been scored against outcomes.
The market has his team at 23.0 points implied.
The index ranks this quarterback and this matchup against every qualifying passer, over games already played. It is not a probability and not a bet — our passing-yards model failed its calibration test, and we publish that on the accuracy page.
Across 6,967 regular-season games since 1999, 24 of 28 spread rungs with a usable sample had a 95% interval containing 50%. In other words the number itself has never been an edge, at any rung, in twenty-seven seasons. What it is good for is knowing which numbers are expensive to cross — and a rung holding fewer than 30 games is shown greyed rather than dropped.