Through four games of 2026, the defending Super Bowl champions are running −0.107 EPA/play on offense — 29th in the league, with only the Chargers, Tampa Bay and Miami worse — and a passing game at −0.144 EPA/play that ranks 26th. Philadelphia is 2–2: wins over Washington (24–22) and Tennessee (24–20), then a 27–7 no-show at Chicago in Week 3 and a 24–20 home loss to the Rams in Week 4.

What's strange about the regression is where it isn't. Jalen Hurts' deep ball and intermediate throws are performing at or above his career-best per-play efficiency. What has broken is the easy stuff — the 0–9 air-yard throws and the behind-the-line screens and dumpoffs that have been the floor of this offense since 2022.

Table of Contents

• The Week 3–4 collapse
• The team-level cliff
• Hurts by target depth
• A straight-line CPOE drift
• What to make of four games

The Week 3–4 collapse

The short-game breakdown isn't a flat four-game story — it bent sharply in the two losses. Philadelphia's team-wide offensive EPA/play by week:

And the same pattern shows up inside Hurts' own short-area passing game (0–9 air yards):

Weeks 1 and 2 were sub-par short-game outings inside a winning offense. Weeks 3 and 4 are a different shape entirely — short-game efficiency that cratered alongside the scoreboard, not independent of it.

The team-level cliff

Zooming out, Philadelphia's offensive EPA/play by season since 2020:

The 2022–2025 window ranges from +0.021 to +0.121 on offense and +0.035 to +0.133 on pass. The 2026 marks aren't just below that range — they're the worst Eagles offensive EPA/play since 2020, on 228 plays, which is a small sample but not a trivial one.

Hurts by target depth

Splitting Hurts' throws by air yards tells a sharply two-sided story. Completion rate, EPA/play, and CPOE (completion percentage over expected) for each of the last four seasons:

Deep (20+ air yards)

Intermediate (10–19 air yards)

Short (0–9 air yards)

Behind the LOS (negative air yards)

The deep (+0.734) and intermediate (+1.187) EPA/play marks are both the best of Hurts' last four seasons. The intermediate CPOE (+9.3) is his best since 2023, and the deep CPOE (+13.0) has no peer in the four-year window. On 35 combined downfield attempts, Philadelphia is getting the version of Hurts the market pays for.

The short game is the opposite story. The 0–9 bucket ran between 70.8% and 74.9% completion across the three prior seasons; it's at 60.3% through four games, with the first negative EPA/play (−0.307) in the sample. Behind-the-line throws — screens, swings, outlet checks — are at 62.5% completion, −0.917 EPA/play and −8.8 CPOE on 16 attempts. That's a small, loud bucket, but its direction matches what the larger 63-attempt short-game sample is doing, which is harder to wave off as noise.

A straight-line CPOE drift

Hurts' overall CPOE, across all attempts, five seasons running:

The 2026 number is a quarter-season on only 105 attempts, but the three-year drift from +7.7 to +2.4 to +1.0 lines up directionally with what the short-game buckets are doing on their own. The downfield shots are compensating in EPA terms; the accuracy baseline that used to anchor the whole operation isn't holding the way it did in 2024.

What to make of four games

Four games is four games. Thirteen deep attempts is thirteen deep attempts. These numbers could move substantially by the bye. But the pattern — downfield passing at or above career-best efficiency, short and behind-the-line passing at the worst marks of the Hurts-starter era — isn't the familiar shape of a Super Bowl hangover, a protection problem or a deep-shot problem. It's the easy-throw layer of the offense that's gone cold, and the week-by-week trace says it got worse, not better, in the two losses.

I'd want another two or three games before calling the short-game numbers a durable regression rather than two ugly outings against Chicago and the Rams. But if Philadelphia is going to climb out of 29th in offensive EPA/play, the fix doesn't look like it needs to come from Hurts throwing deeper — it needs to come from the throws that used to be automatic going back to being automatic.