> ## Content Index > Fetch the complete content index at: https://www.statsdigest.com/llms.txt > Use this file to discover other available public pages before exploring further. # Kansas City's Offense Through Three Weeks Looks Like 2018-2022, Not 2023-2025 - URL: https://www.statsdigest.com/chiefs-offense-2026-looks-like-peak-mahomes/ - Published: 2026-10-03T11:13:38.000Z - Updated: 2026-10-03T11:13:37.000Z - Description: Through three games, Kansas City's EPA per play looks like the Mahomes of 2018-2022, not 2023-2025. - Author: Bill Seifert - Tags: Kansas City Chiefs, Patrick Mahomes, EPA, NFL analytics, offense, quarterback play The Chiefs won two Super Bowls between 2023 and 2025 with an offense that, by expected points added, was merely fine. Patrick Mahomes posted 0.07 EPA per dropback in 2023 and roughly 0.10 in 2024 — not anything you associated with Mahomes before 2023. Through three games of 2026, that number is 0.236. The sample is three games and 191 offensive plays. That is the first and most important thing on this page. But the gap between what Kansas City is doing right now and what Kansas City did for three straight seasons is large enough, and consistent enough across both pass and run, that it is worth putting on the record before week 4 — because if it holds even halfway, the story of the Mahomes era is about to need a new chapter. ## Table of Contents 1. The number that changed 2. The 2023–2025 baseline 3. Mahomes himself 4. The run game, which is also back 5. What the sample actually supports ## The number that changed Through three games, Kansas City's offense is averaging **+0.184 EPA per play** on 191 pass and run snaps — fourth-best in the league, behind only San Francisco (+0.357, in two games), Buffalo (+0.218), and Baltimore (+0.193). The pass game is averaging **+0.236 EPA per dropback**; the run game, **+0.125 EPA per carry**, second in the NFL behind only Buffalo's +0.178. The games themselves: a 31–10 win over Denver in week 1 (+0.164 EPA/play, with the run game carrying it at +0.366), a 33–30 win over Indianapolis in week 2 (+0.229 EPA/play, +0.274 passing), and a 24–10 win at Miami in week 3 (+0.138 overall, with Mahomes posting an eye-popping **+0.539 EPA per dropback** on 24 passes while the run game went backward at −0.246). Three different flavors of good offense in three weeks. The composite is a team success rate of **50.3%**, up from 44.5% in 2025. ## The 2023–2025 baseline Here is why that composite matters. Kansas City's offensive EPA per play, by season: | Season | Overall EPA/play | Pass EPA/db | Run EPA/carry | | --------------- | ---------------- | ----------- | ------------- | | 2018 | +0.227 | +0.311 | +0.088 | | 2019 | +0.156 | +0.230 | +0.030 | | 2020 | +0.182 | +0.258 | +0.048 | | 2021 | +0.152 | +0.199 | +0.070 | | 2022 | +0.187 | +0.255 | +0.074 | | **2023** | **+0.028** | **+0.050** | **−0.008** | | **2024** | **+0.062** | **+0.082** | **+0.031** | | **2025** | **+0.035** | **+0.040** | **+0.027** | | **2026 (3 gm)** | **+0.184** | **+0.236** | **+0.125** | From 2018 through 2022, KC's pass EPA per dropback never dropped below 0.199\. From 2023 through 2025, it never got above 0.082\. In 2025, Kansas City's offensive EPA per play was +0.035 — a perfectly good offense. Not the one that ran anywhere from +0.152 to +0.227 EPA per play for five straight seasons before 2023. The 2026 pass number (+0.236) sits inside the 2018–2022 band. The 2026 run number (+0.125) is better than every KC season back to 2018\. The overall figure (+0.184) is a stride ahead of the 2023–2025 average of roughly +0.042. ## Mahomes himself Mahomes personally, through 102 dropbacks: - 67-of-102, **812 yards**, 7 TD, 2 INT - **+0.236 EPA per dropback** - **+3.21 CPOE** For the 2018–2022 Mahomes — the one that won an MVP and a Super Bowl and reset what quarterback play looked like — the aggregate across 3,636 dropbacks was **+0.256 EPA per dropback with +3.14 CPOE**. It's worth saying plainly: stacking a 102-dropback sample against a 3,636-dropback career baseline is not an apples-to-apples confidence comparison — the 2026 number carries a far wider error bar, and three more bad quarters could move it a lot. But the 2026 figure currently sits inside that era's range rather than anywhere near the 2023–2025 numbers, which is informative even with that caveat attached. Compare against the three-year dip. Mahomes' EPA per dropback: 0.069 in 2023, 0.096 in 2024, 0.127 in 2025 — a slow climb, but his CPOE went the other way (3.87 → 2.30 → 0.34). The 2026 CPOE snapping back to +3.21 is the piece of this that feels least like noise, because CPOE tends to stabilize faster than EPA does. He is throwing more accurate passes than the model expects, on a scale he hadn't hit since the title years. ## The run game, which is also back The quieter part of the story. The lead back — logged in the play-by-play as K. Walker — has 65 carries for **360 yards** and 2 TDs at **+0.092 EPA per carry**. The backup, E. Johnson, has added 16 carries for 56 yards at −0.013\. Mahomes himself has 7 carries for 45 yards and a touchdown at +0.440\. In aggregate, the Chiefs are running the ball at +0.125 EPA per carry — second in the league behind only Buffalo's +0.178, and better than any KC season dating back to 2018. Travis Kelce, Rashee Rice, Xavier Worthy, and Tyquan Thornton have all factored into the passing game early, though a confirmed target-by-target EPA breakdown is a piece this article doesn't have the sample depth to responsibly state as fact yet. ## What the sample actually supports Three games. 191 plays. The 102 dropbacks are enough to put real weight on Mahomes' CPOE but not enough to settle an EPA number — and even the CPOE read carries a wide error bar against the 3,636-dropback career aggregate it's being measured next to. The schedule hasn't been a gauntlet — Denver, Indianapolis, Miami — and the Buffalo, Baltimore, and divisional tests are still ahead. What I'd say the data supports, plainly: - The 2026 numbers through three games are a departure from the 2023–2025 baseline, not a continuation of it. The gap (overall offense roughly +0.14 EPA/play better, pass about +0.17 better, run about +0.10 better) is too wide to wave off as noise even at 191 plays. - The CPOE jump from +0.34 in 2025 to +3.21 in 2026 is the most believable piece of early evidence, because CPOE tends to stabilize faster than EPA. - The run game — a workhorse back, Mahomes' own scrambling, a +0.125 EPA/carry that already beats every KC season since 2018 — is the clearest non-QB-dependent piece of the early turnaround. - What this piece is **not** saying: that KC's offense is finished as a league-average unit for good, that Mahomes "fixed" something specific, or that we know how this looks after Buffalo in week 12\. I'd want to see this through at least six games before I'd call it settled. But if you were waiting for a sign that the offense that disappeared after Super Bowl LVII was coming back, this is what it would look like through three games.