Defense rankings — 2026 Week 8

Use this for the order, not the number. Ranking defenses this way is about twice as good as ranking them by season-to-date fantasy points (Spearman 0.29 out of sample vs 0.13) — but half of what a defense scores is turnovers and touchdowns, which nothing predicts.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Defense ProjFloor–ceiling Pts allowedfrom tiers SacksTurnovers
1 PIT vs CLE 8.4 2.1 – 15.6 17.1 +1.9 2.9 1.4
2 DAL vs ARI 8.1 1.8 – 15.2 18.6 +1.5 2.8 1.5
3 GB vs CAR 7.6 1.3 – 14.7 19.5 +1.3 2.7 1.4
4 NYJ vs LV 7.1 0.8 – 14.3 19.9 +1.1 2.6 1.3
5 LV @ NYJ 6.7 0.4 – 13.8 21.4 +0.8 2.5 1.3
6 CLE @ PIT 5.8 -0.5 – 13.0 23.2 +0.3 2.4 1.2
7 BUF vs BAL 5.6 -0.7 – 12.7 24.7 -0.1 2.2 1.3
8 BAL @ BUF 4.8 -1.5 – 11.9 27.2 -0.7 2.1 1.3
9 CAR @ GB 4.7 -1.6 – 11.8 26.7 -0.5 2.1 1.2
10 ARI @ DAL 3.7 -2.6 – 10.8 29.5 -1.1 1.8 1.1
Projected points allowed comes from the posted total and the opponent's implied total; the tier column is the expectation across the points-allowed tiers, not the tier a single projection happens to land in. Defensive touchdowns and safeties are the league mean — nothing forecasts them (fitted 2019-2025). Measured walk-forward, one season held out at a time: Spearman 0.295, MAE 4.27 against a 4.41 league-mean baseline, R² 0.079 over 3,136 defense-games. The R² is the honest ceiling: this orders defenses, it does not predict their scores. Study at backtests/kicker_dst/FINDINGS.md.