Week 5 in review

Flattered and undersold

Flattered · the score ran ahead of the play by more than 7 points

• Texas Tech at Colorado, 29–7 W+22 vs -0 implied · +22
  The scoreboard flattered Texas Tech: the +22 margin ran 22 points ahead of the -0 implied by per-play quality (PPA (expected points added per play) -0.22 vs -0.18, success rate 30% vs 40%) (Colorado turned it over 4 times).
• Houston vs UCF, 27–17 W+10 vs -4 implied · +14
  The scoreboard flattered Houston: the +10 margin ran 14 points ahead of the -4 implied by per-play quality (PPA (expected points added per play) +0.03 vs +0.12, though Houston was more consistent: success rate 51% vs 30%) (UCF turned it over once).
• Louisville at NC State, 28–31 L-3 vs -12 implied · +9
  The scoreboard flattered Louisville: the -3 margin ran 9 points ahead of the -12 implied by per-play quality (PPA (expected points added per play) +0.24 vs +0.30, though Louisville was more consistent: success rate 48% vs 40%) (NC State turned it over twice).
• SMU vs Boston College, 25–16 W+9 vs +1 implied · +8
  The scoreboard flattered SMU: the +9 margin ran 8 points ahead of the +1 implied by per-play quality (PPA (expected points added per play) +0.08 vs +0.06, success rate 36% vs 40%) (Boston College turned it over once).

Undersold · the score ran behind the play by more than 7 points

• Florida State vs Virginia, 38–7 W+31 vs +43 implied · -12
  The scoreboard undersold Florida State: the +31 margin ran 12 points behind the +43 implied by per-play quality (PPA (expected points added per play) +0.33 vs -0.29, success rate 50% vs 29%) (Florida State turned it over once).
• Alabama at Mississippi State, 56–23 W+33 vs +40 implied · -7
  The scoreboard undersold Alabama: the +33 margin ran 7 points behind the +40 implied by per-play quality (PPA (expected points added per play) +0.46 vs -0.16, success rate 60% vs 35%).
• Kentucky at South Carolina, 35–34 W+1 vs +8 implied · -7
  The scoreboard undersold Kentucky: the +1 margin ran 7 points behind the +8 implied by per-play quality (PPA (expected points added per play) +0.19 vs +0.09, success rate 50% vs 49%).

Closest projection

Over the 20 games this week where SP+, Elo and the line all existed before kickoff, the line as a benchmark (its miss only, never closest) (2 more shown on the Projections page but not in the record: no page before kickoff).

Our model experimental · 21 eligible games · vs SP+: 11.7 against 10.9, closest 8 / 12 · vs Elo: 11.7 against 12.3, closest 11 / 10 · vs Line: 11.7 against 11.2, closest 8 / 12 · all four: our model closest 5 of 20

Our call · 20 games where SP+, Elo and Our model all existed before kickoff, each model on those same games, and the line for comparison (its miss only: a benchmark is never counted as naming the winner):

Against its benchmark, the line: Our call was at least as close in 7 of 20, the line in 13 (a tie counts for both).

Upsets

Ranked teams that lost to an unranked or lower-ranked opponent, by the AP Top 25 in effect for week 5, with each model's pregame margin from the ranked side and the line beside them for comparison only; ✓ when a model had the winner, ✗ when it did not.

• #8 Florida lost at #25 MissouriMissouri 45, Florida 17
  SP+ ✗

  +2.3

  Elo ✗

  +0.4

  Line

  +5.5

  Our model ✗

  +0.8

Called it: SP+ 0 of 1 · Elo 0 of 1 · Our model 0 of 1.

Biggest swings

The plays that moved win probability most, across the 22 games with win-probability data, from the home side.

• Texas Tech at ColoradoColorado 0% → 47%
  Q4 0:08, 1st & 10: (00:08) No Huddle-Shotgun #28 A.Rowden rush middle for 8 yards gain to the TTU12 (#99 D.Manuel; #17 L.Lefau) · final Colorado 7, Texas Tech 29
• Pittsburgh at Virginia TechVirginia Tech 21% → 62%
  Q4 10:13, 2nd & 8: (10:13) Shotgun #17 E.Grunkemeyer pass complete deep right to #11 D.Brown caught at UP 02, for 28 yards to the UP 00 TOUCHDOWN, clock 10:07, 1ST DOWN #17 J.Love kick attempt good (H: #95 N.Totten, LS: #96 C.Epling) · final Virginia Tech 33, Pittsburgh 35
• Kentucky at South CarolinaSouth Carolina 65% → 28%
  Q3 0:47, 1st & 10: (00:47) Shotgun #3 K.Minchey pass complete deep right to #4 N.Anderson caught at USC00, for 32 yards to the USC00 TOUCHDOWN, clock 00:40, 1ST DOWN #88 A.Zouagui kick attempt good (H: #93 W.Berry, LS: #47 S.Radnoti) · final South Carolina 34, Kentucky 35

Regression watch, checked

Last week's teams whose score ran more than 7 points from the play, and what they did this week against their season-to-date form (average margin before the game, home field removed). The 2022–25 study found flattered teams fell -3.0 short of form the next game and undersold teams beat it by +1.7.

• Georgia · flattered +38 vs Oklahomavs Vanderbilt 38–14 W · -18.8 vs form (-8.7 vs the field)
• Miami · flattered +17 vs Central Michiganat Clemson 41–13 W · -12.2 vs form (-2.2 vs the field)
• Notre Dame · flattered +11 vs Purdueat North Carolina 37–26 W · -19.9 vs form (-9.8 vs the field)
• Iowa · flattered +10 vs Michiganvs Ohio State 14–31 L · -43.0 vs form (-32.9 vs the field)
• Florida · flattered +9 vs Ole Missat Missouri 17–45 L · -55.0 vs form (-44.9 vs the field)
• Houston · flattered +8 vs Georgia Southernvs UCF 27–17 W · -16.5 vs form (-6.4 vs the field)
• UCLA · flattered +7 vs Marylanddid not play
• Oklahoma · undersold -38 vs Georgiadid not play
• Oklahoma State · undersold -16 vs West Virginiadid not play
• SMU · undersold -13 vs Missouri Statevs Boston College 25–16 W · -5.8 vs form (+4.3 vs the field)
• Michigan · undersold -10 vs Iowaat Minnesota 14–20 L · -11.5 vs form (-1.4 vs the field)
• Texas Tech · undersold -9 vs Sam Houstonat Colorado 29–7 W · +8.0 vs form (+18.1 vs the field)
• Ole Miss · undersold -9 vs Floridadid not play
• Boise State · undersold -7 vs Western Michiganvs Utah State 37–18 W · +4.0 vs form (+14.1 vs the field)

Baseline: all 29 page teams that played ran -10.1 vs form this week. The baseline matters because early-season form comes mostly from easier non-conference games, so the whole field tends to fall short of it once conference play starts; a flagged group is read against the field, not against zero.

Flattered: -27.6 vs form over 6, -17.5 against the baseline (-3.0 in the study) · Undersold: -1.3 vs form over 4, +8.8 against the baseline (+1.7 in the study). One week is a few teams and proves nothing either way; the point is that the check is visible every week.