KC
Kansas City Chiefs
21
AWAY
—
BUF
Buffalo Bills
30
HOME_WIN
BUF TOTAL EPA
+14.2
on offense
KC TOTAL EPA
-8.6
on offense
PROFESSOR'S OPENING: The score says 9 points. The EPA says 22.8 points of
offensive advantage. The box score makes this look competitive. It wasn't, by the numbers.
Kansas City's +0.055 EPA/play profile didn't show up. Buffalo's +0.141 profile did.
Three turnovers worth of expected points is hard to overcome even for Mahomes.
DRIVE LOG — ALL 18 POSSESSIONS
Starred drives (★) were decisive EPA swings — defined as >4.0 WP shift on a single play.
Open full game in
DRIVE_TRACKER.
1
KC
★ Mahomes INT on play 2 — Taron Rapp pick, roughing penalty offset
TURNOVER
-3.1
-6.8%
2
BUF
Allen: 3rd-and-6 to Cooper +30, 3rd-and-8 to Samuel +15, J.Cook TD
TOUCHDOWN
+5.4
+11.9%
3
KC
Mahomes sacked on 3rd-and-5, punt
PUNT
-2.8
-6.7%
4
BUF
Allen sustains 10-play drive, converted 3rd-and-2, Allen keeper TD
TOUCHDOWN
+6.1
+14.2%
5
KC
Rice +20, Mahomes finds Travis Kelce in red zone — 1st KC score
TOUCHDOWN
+4.2
+8.3%
6
BUF
Allen sacked, 3-and-out
PUNT
-2.2
-4.1%
7
KC
3rd-and-6 conversion, Mahomes scramble sets up FG
FIELD_GOAL
+1.8
+3.4%
8
BUF
Shakir over the middle, Diggs slot for 18, Dalton Kincaid TD
TOUCHDOWN
+4.9
+10.7%
9
KC
★ Mahomes fumble, BUF recovers — 18-point deficit at half
TURNOVER
-5.6
-12.3%
10
BUF
Converted turnover, Allen bootleg TD — 30-10 at one point
TOUCHDOWN
+3.8
+9.1%
11
KC
Mahomes starts Q3 comeback — Kelce TD from 22 yards
TOUCHDOWN
+4.7
+9.5%
12
BUF
Allen kneels twice — clock management, BUF protecting lead
PUNT
-0.8
-1.2%
13
KC
Pacheco for 11, incomplete, Mahomes sacked for -8 — another punt
PUNT
-3.1
-5.8%
14
BUF
Allen drains 7 plays / 4:22 off clock — BUF field goal
FIELD_GOAL
+1.3
+3.2%
15
KC
KC FG drive — Mahomes to Rice, settles for short field goal
FIELD_GOAL
+1.1
+2.2%
16
BUF
Allen runs clock — four runs, BUF goes 3-and-out intentionally
PUNT
-1.1
-0.9%
17
KC
★ Mahomes scores — too little, needs onside. 30-21 final swing
TOUCHDOWN
+3.3
+4.9%
18
BUF
Onside recovery — Allen kneels. Game over.
KNEEL
-0.2
-96.0%
THE STORY IN THREE DRIVES
DRIVES 1–2 (KC TURNOVER → BUF TD).
Mahomes' second-play interception followed immediately by a 65-yard Buffalo touchdown drive
was a 18-point expected-value swing before the first commercial break. The roughing penalty
that offset the interception — and gave Kansas City a first down — would have mattered had
KC scored. They didn't. The Bills went the length of the field on their next possession and
never looked back.
DRIVE 9 (MAHOMES FUMBLE, END OF HALF).
KC was already losing EPA on the night but still within two scores at 24-10. The fumble
going into halftime turned a 14-point game into one where KC would need to be perfect in
the second half. Against a Bills defense playing at home, that wasn't happening.
DRIVES 13–14 (KC SACK, BUF CLOCK KILL).
When Kansas City needed drives in the third quarter, they generated another sack-punt.
Buffalo followed it with a 4-minute field goal drive. At that point the win probability
model had KC at roughly 8%. The comeback path was closed.
WHAT THE DATA SAID BEFORE AND WHAT IT SAYS NOW
The pre-game EPA gap (BUF +0.141, KC +0.055) projected Buffalo as the better offense by
roughly a touchdown over the course of a game if both teams ran true-to-form. They did.
Buffalo ran 50 offensive plays at their season EPA average. Kansas City ran 47 plays and
generated negative total EPA — well below their season average.
TAKEAWAY: Two KC turnovers generated a combined -8.7 expected points.
Remove those plays and this is a different game. But turnovers aren't random — they correlate
with pressure, and BUF's pass rush forced Mahomes into uncomfortable situations all night.
The defense created what the offense cashed in on. That's how you beat Mahomes.
PROFESSOR'S VERDICT
Buffalo was the better team by EPA in 2024 through Week 11. This game reflected that.
The final score of 30-21 undersells the margin; the expected-points story had BUF ahead
by ~22.8 points in EPA, which the actual deficit only partially captures because KC
scored late against a prevent defense. File this under: the EPA model was right.