Through three games of 2026, opposing offenses have averaged +0.398 EPA per dropback against the Cardinals across 84 dropbacks (pass attempts plus sacks). That same number was +0.063 across 581 dropbacks in 2024 and +0.150 across 603 dropbacks in 2025. Different sample sizes, obviously — three games is a fraction of a season — but the direction isn't subtle, and the mechanism is weirder than "the deep ball blew up."

Table of Contents

• Where the leak is
• What the small sample does and doesn't say

Where the leak is

Break 2026 down by target depth and the biggest swing from Arizona's own two-year baseline isn't on 20-yard bombs — it's on throws at or inside 10 air yards.

2026, three games:

• Short (0–10 air yards): 34 attempts, 24 completions, +0.702 EPA/play, 23.9 total EPA
• Intermediate (11–20 air yards): 29 attempts, +0.343 EPA/play, 10.0 total EPA
• Deep (21+ air yards): 8 attempts, +0.837 EPA/play, 6.7 total EPA
• Behind the line of scrimmage: 9 attempts, −0.126 EPA/play
• Sacks: 4, −1.49 EPA each, −6.0 total

Compare that short-pass number to where Arizona lived for two full seasons under Jonathan Gannon: +0.167 EPA/play on 304 short attempts in 2024, +0.198 EPA/play on 323 attempts in 2025 (218 and 233 completions, respectively — right around a 72% completion rate both years). Through three games of 2026 the completion rate is similar, 24 of 34 (70.6%), but the EPA damage per play has more than tripled.

That shows up in the yardage, too. Yards per completion on short throws have gone from 8.48 (2024) and 8.12 (2025) to 13.08 so far in 2026. And the share of those completions going for 20-plus yards — a rough proxy for "turned into a chunk play after the catch" — has moved from 4.6% (2024) and 3.0% (2025) to 8.3% in 2026. On a 24-completion sample, that last swing is roughly two plays; call it a flag, not a trend.

The other buckets don't tell the same story. The deep-ball number is the loudest on paper — +0.837 EPA/play — but it's on just 8 attempts, and Arizona's deep coverage was already shaky in 2025 (+0.722 on 54 attempts), so this isn't new, it's a continuation on a tiny sample. The intermediate bucket is actually better than either of the last two years (+0.343 in 2026 vs. +0.372 in 2025 and +0.583 in 2024). And the pass rush — the usual first culprit when a defense slips — is in the same neighborhood as last year or better: a 4.8% sack rate on 84 dropbacks in 2026, against 5.0% in 2025 and 7.1% in 2024.

Put together: the deep ball was already a known issue, the middle of the field has quietly improved, and the pass rush hasn't collapsed. The one number that's moved clearly outside its normal range is the short stuff — exactly the area a defense is supposed to control with leverage and tackling, not scheme exotics.

What the small sample does and doesn't say

Three games, 34 short attempts, 24 short completions. A dropped slant or a missed tackle in week 4 will move these rates more than a quarter-season of normal football would. This is a flag, not a trend, and nobody should be writing Arizona's defensive obituary off of it.

But it's a specific flag with a specific shape. Short-pass EPA allowed sat in the +0.17-to-+0.20 range for two full seasons under this coaching staff, and it's at +0.70 through three games — a jump the deep-ball and pass-rush numbers can't explain, because those are either normal or already a known issue. If this holds up, it reads less like a scheme failure and more like something that shows up on film as missed tackles and soft leverage underneath — fixable game-to-game, not a four-year rebuild. Whether that's the real story or just three weeks of bad tackling luck is the thing worth watching before anyone calls it a trend.