Week 9 key numbers: where the model and the market disagree across 3 and 7

Some numbers matter more than others on a football spread. A field goal is worth 3 points, and a touchdown with the extra point is worth 7, so games land on 3 and 7 more often than on any other margin. That is what a key number is: a margin the final score reaches often enough that the half point on either side of it changes what a spread is worth. Every game here has two numbers, the model's margin and the market's spread, from the site feed on November 5, rounded to the nearest half point.

Why 3 and 7 land more often

The model prices every margin, and the price shows why 3 and 7 stand out. In the Denver at Carolina game, the model gives a 7.5% chance of landing exactly on 3, and a 4.5% chance of landing exactly on 7. Every other whole margin from 1 to 10 sits between 1.5% and 4%. A spread of 3.5 asks the favorite to win by more than a field goal. A spread of 3 asks for exactly a field goal or better, with a push if it lands on the number.

Week 9, model against market

Across the fifteen Week 9 games, the model and the market agree on the favorite in all fifteen. They disagree on the size. Eight games disagree by 2.5 points or more. One game, Houston at the Chargers, has the two numbers exactly together at 4.5. The chart below shows every game, model number in gold and market spread in grey, ordered by the size of the gap. The dashed lines mark 3 and 7.

The widest gaps

Tampa Bay at Chicago carries the widest gap of the week. Chicago is a 6.5 point favorite in the market and a 2 point favorite in the model, a gap of 4.5 points, and the card's spread call is Tampa Bay plus 6.5, in the moderate tier. The model builds Chicago's 2 points from home field at 0.5, offense at 2.5, defense at minus 0.5, quarterbacks at minus 0.5, and other factors at 0. The market's number sits a hook below 7, and the model's number sits below 3.

Denver at Carolina runs the other way: Denver is a 2.5 point favorite in the market and a 6 point favorite in the model, a gap of 3.5 points, with Denver minus 2.5 in the moderate tier. Dallas at Indianapolis has the same shape: Dallas is a 5.5 point favorite in the market and a 2.5 point favorite in the model, a gap of 3 points, with Indianapolis plus 5.5 in the moderate tier.

Cleveland at New Orleans is the week's moving line. New Orleans is a 3.5 point favorite in the market and a 1 point favorite in the model, a gap of 2.5 points, with Cleveland plus 3.5 in the slight tier. The spread moved from 2.5 down to 1.5 on November 3 and back up to 3.5 on November 4. Jacksonville at Baltimore opens the week: Jacksonville is a 9.5 point favorite in the market and a 7 point favorite in the model, a gap of 2.5 points, with Baltimore plus 9.5 in the slight tier. Baltimore starts Joe Fagnano at quarterback with Lamar Jackson out and Tyler Huntley out. The total tells the other half: the model total is 46, the market total is 39.5, and the total call in the feed is Over 39.5, in the strong tier.

Three games sitting on 7

The model's number is exactly 7 in three games, and in all three the market sits at 9.5. Jacksonville at Baltimore is one. The others are the Jets at Kansas City and Las Vegas at San Francisco. In each, the model has the favorite by a touchdown, and the market has the favorite by a touchdown and a field goal. The chance of landing on 7 is about 4.5%. Detroit at Miami sits a hook below 7 on both numbers, the model at 5.5 and the market at 6.5. Arizona at Seattle runs the other way, the model at 10.5 and the market at 8.5. Six games carry a market spread of 6.5 or more, and in five of them the model sits below the market number.

The line that moved on a quarterback

Philadelphia is the week's clearest example of a number moving on news. The Giants are a 3.5 point favorite in the market, and the model has the Giants by 0.5, a gap of 3 points. The spread has held at 3.5 since November 3. The model moved on November 5, from the Giants by 1 to the Giants by 0.5, when Philadelphia's starter changed from Tanner McKee to Joe Milton III. The moderate moneyline call on Philadelphia was removed the same morning. The total moved too, from 40.5 to 39.5 on November 3 and to 38.5 on November 4. The spread stayed still while everything around it moved.

What the season has shown so far

Through Week 8, the graded NFL card sits at 76-80-1, +0.11u. The one push in that record is a moneyline on the October 11 tie between the Giants and Washington, not a spread landing on 3 or 7. No graded spread lean has pushed on a key number yet this season. A 7.5% chance on 3 and a 4.5% chance on 7 means most weeks pass without a push, and then one lands and the half point decides the result. Every spread here is the market number as of November 5. The comparison that counts for grading is the model's number against the market's number at the lock, two hours before kickoff. Read the gap, then read the lock.

21+. If gambling stops being fun, call or text 1-800-GAMBLER.