Week 1 Recap: What Held, What Didn't, and What Changes in Week 2

• Touchdown scorers. Every 90+ anytime TD score hit: Gibbs, Hampton, Henry, Taylor, and Walker and Rice on Monday night. The top ten by score went 8-2 and the top twenty went 13-7.
• Primetime. Opening night went 4-2 and Monday night went 7-3. Primetime as a whole finished 13-11 and positive at served prices.
• Passing yards. Yardage unders went 10-9 at 90+ and 18-14 overall, in a week when quarterbacks threw for 4.8% more than their lines.

• The 1pm window. Eight games, 458 points, 25 passing touchdowns, and a 96-point Bears-Panthers game that is now the highest-scoring opener ever. Our 1pm 90+ scores went 17-26. Nearly the whole week's loss sits in those three hours.
• Passing touchdowns. Half of the starting quarterbacks threw two or more touchdown passes. In the last three openers, fewer than a third did. Under 1.5 passing TDs went 7-10 at 90+. While yardage held, TDs and attempts did not.

League scoring fell from 24.8 points per team in 2020 to 21.8 in 2023 as defenses moved to two-high safety shells, then crept back to 23.0 last season. Week 1 came in at 24.7 per team, 2020 territory. One week proves nothing, but if offenses have answered the two-high look, the touchdown environment will not look like 2022 through 2025, and the model reads that through the lines and results as they arrive.

The in-season model takes over Thursday. Week 1 is the noisiest week of the year by default: new rosters, new schemes and one game of data per team, and this year added a historic scoring window on top. From here the samples stabilize, the inputs settle and the scores with them. Week 2 is exactly the environment this model was built for: less noise, more data, steadier scores, and sixteen more games to show it. Our models will only continue to improve as more relevant data comes in each week.