ROBOT
POLL
WEEK 14 · 2025 SEASON · RENDERED FROM LOCKED MODEL · NEVER EDITED
336–335–14
VS THE CLOSING LINE · 50.1%
BRIER .190 CALIBRATED ✓
A pundit with a falsifiable public record. Every pick comes with a receipt. Every receipt gets graded. The losses stay up.
THE POLL PICKS DISAGREES BETS 💰 SIGNAL MINE SCOREBOARD AUTOPSY HISTORY LAB THE ROBOT 🤖
🗳 THE ROBOT POLL — TOP 25
ONE VOTER · NO BIAS · NO BRAND LOYALTY · NO MERCY
Least-squares ratings from actual results: rating = points better than the average FBS team on a neutral field. Margins cap at 28 — style points don't count past four touchdowns. Strength of schedule falls out of the math, not a committee's mood.
1
Ohio State 10–0 —
SOS #45 · best win: Washington
+26.4
2
Texas Tech 9–1 —
SOS #55 · best win: Utah · worst loss: Arizona State
+24.9
3
Indiana 10–0 ▲1
SOS #32 · best win: Oregon
+23.7
4
Utah 8–2 ▼1
SOS #28 · best win: Arizona State · worst loss: BYU
+22.5
5
Notre Dame 9–2 —
SOS #37 · best win: USC · worst loss: Miami
+22.3
6
Texas A&M 10–0 ▲1
SOS #23 · best win: Notre Dame
+19.4
7
Oregon 9–1 ▲3
SOS #46 · best win: USC · worst loss: Indiana
+19.2
8
BYU 9–1 —
SOS #7 · best win: Utah · worst loss: Texas Tech
+18.9
9
Miami 8–2 —
SOS #36 · best win: Notre Dame · worst loss: Louisville
+18.3
10
Georgia 9–1 ▼4
SOS #38 · best win: Ole Miss · worst loss: Alabama
+18.2
11
Alabama 8–2 —
SOS #11 · best win: Georgia · worst loss: Florida State
+16.2
12
Vanderbilt 8–2 ▲3
SOS #49 · best win: Missouri · worst loss: Texas
+15.6
13
Oklahoma 8–2 —
SOS #19 · best win: Alabama · worst loss: Texas
+15.3
14
USC 8–3 ▼2
SOS #35 · best win: Iowa · worst loss: Illinois
+14.4
15
Washington 7–3 ▲3
SOS #44 · best win: Illinois · worst loss: Wisconsin
+14.2
16
Ole Miss 9–1 ▼2
SOS #34 · best win: Oklahoma · worst loss: Georgia
+14.2
17
South Florida 7–3 ▲2
SOS #72 · best win: North Texas · worst loss: Navy
+13.1
18
Tennessee 7–3 ▲2
SOS #56 · best win: Arkansas · worst loss: Oklahoma
+12.8
19
SMU 7–3 ▲5
SOS #61 · best win: Miami · worst loss: Wake Forest
+12.6
20
Arizona 7–3 ▲1
SOS #41 · best win: Kansas State · worst loss: Houston
+12.5
21
Iowa 6–4 ▼5
SOS #40 · best win: Penn State · worst loss: Iowa State
+12.3
22
Michigan 9–2 NEW
SOS #52 · best win: Washington · worst loss: USC
+11.9
23
Penn State 4–6 NEW
SOS #27 · best win: Nebraska · worst loss: UCLA
+11.6
24
Texas 8–3 ▼2
SOS #42 · best win: Vanderbilt · worst loss: Florida
+11.2
25
Illinois 6–4 ▼8
SOS #10 · best win: USC · worst loss: Wisconsin
+10.7
🧾 Does the poll mean anything? Over 442 ranked-vs-ranked games since 2016, the higher-ranked team won 63.1%. Teams ranked within 5 spots of each other split nearly even (54.7%) — if the poll says #7 vs #9 is a coin flip, believe the poll, not the seeding.
🤖 THE ROBOT DISAGREES — THE 3 GAMES WHERE VEGAS AND I ARE NOT FRIENDS
JACKSONVILLE STATE +1.5
+25 PTS APART
VEGAS SAYS
46%
ROBOT SAYS
71%
The receipt's loudest clue is rating_gap at +0.4 points of lean. Vegas and I are 25 percentage points apart on this game. One of us is wrong, and it's on the record either way.
TROY +7.0+20 PTS APART
Vegas 33% · Robot 53%
The receipt's loudest clue is rating_gap at -0.3 points of lean. Vegas and I are 20 percentage points apart on this game. One of us is wrong, and it's on the record either way.
NORTHERN ILLINOIS -5.0+20 PTS APART
Vegas 62% · Robot 83%
The receipt's loudest clue is rating_gap at +1.2 points of lean. Vegas and I are 20 percentage points apart on this game. One of us is wrong, and it's on the record either way.
💰 THE BETTING ENGINE
$100 $250 $500 $1000
Scans every FBS game with a posted line. Edge = robot's cover probability minus the vig-implied break-even (52.4%). Sized by quarter-Kelly, capped at 20% of budget. Receipts attached to everything.
BET $12 · Ole Miss -7.0 edge +23.7
robot covers 76.1% · market needs 52.4% · receipt ▸
BET $11 · Jacksonville State +1.5 edge +21.0
robot covers 73.3% · market needs 52.4% · receipt ▸
BET $10 · Northern Illinois -5.0 edge +20.8
robot covers 73.2% · market needs 52.4% · receipt ▸
BET $9 · Memphis -3.5 edge +18.5
robot covers 70.9% · market needs 52.4% · receipt ▸
BET $9 · Troy +7.0 edge +17.6
robot covers 69.9% · market needs 52.4% · receipt ▸
BET $8 · Liberty +2.5 edge +15.6
robot covers 67.9% · market needs 52.4% · receipt ▸
BET $7 · North Texas -20.0 edge +14.7
robot covers 67.0% · market needs 52.4% · receipt ▸
BET $7 · Penn State -14.0 edge +14.6
robot covers 67.0% · market needs 52.4% · receipt ▸
BET $7 · UAB +9.0 edge +14.4
robot covers 66.8% · market needs 52.4% · receipt ▸
BET $7 · Eastern Michigan +10.0 edge +14.1
robot covers 66.4% · market needs 52.4% · receipt ▸
everything else: no edge, no bet $13 STAYS IN YOUR POCKET
The robot bets fake money. What you do with real money is a conversation between you and your group chat.
THE WEEK 14 SLATE · TOP 40 BY COMBINED ELO
#1 Ohio State @ Michigan
Sat Nov 29 · line: Michigan +9.5 · ATS: Ohio State -9.5
76%
OHIO STATE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+12.9
rest_diff
+0.0
opp_pass_def_trend
−0.0
rush_def_trend
+0.0
pass_off_trend
−1.2
lookahead_spot
+1.7
travel_km
−0.2
baseline
−0.5
LEAN +10.6→ 76% OHIO STATE
#2 Oregon @ Washington
Sat Nov 29 · line: Washington +7.0 · ATS: Oregon -7.0
74%
OREGON ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+11.0
rest_diff
+0.0
opp_pass_def_trend
+0.2
rush_def_trend
+0.1
pass_off_trend
−0.2
lookahead_spot
+1.5
travel_km
−0.4
baseline
−0.5
LEAN +9.8→ 74% OREGON
#3 Texas A&M @ Texas
Sat Nov 29 · line: Texas +2.5 · ATS: Texas +2.5
54%
TEXAS ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
−0.6
rest_diff
−0.0
opp_pass_def_trend
+0.1
rush_def_trend
−0.7
pass_off_trend
+0.4
lookahead_spot
−0.4
travel_km
+0.1
baseline
+0.5
LEAN +1.4→ 54% TEXAS
#4 Vanderbilt @ Tennessee
Sat Nov 29 · line: Tennessee -2.0 · ATS: Tennessee -2.0
56%
TENNESSEE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+1.1
rest_diff
−0.0
opp_pass_def_trend
−0.1
rush_def_trend
+0.2
pass_off_trend
−1.1
lookahead_spot
−0.7
travel_km
+0.2
baseline
+0.5
LEAN +2.1→ 56% TENNESSEE
#5 Alabama @ Auburn
Sun Nov 30 · line: Auburn +7.0 · ATS: Alabama -7.0
74%
ALABAMA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+11.7
rest_diff
0.0
opp_pass_def_trend
−0.3
rush_def_trend
+0.7
pass_off_trend
−1.0
lookahead_spot
+1.1
travel_km
−0.2
baseline
−0.5
LEAN +9.6→ 74% ALABAMA
#6 Georgia @ Georgia Tech
Fri Nov 28 · line: Georgia Tech +16.5 · ATS: Georgia -16.5 · neutral site
90%
GEORGIA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
0.0
rating_gap
+18.4
rest_diff
+0.0
opp_pass_def_trend
−0.2
rush_def_trend
+0.4
pass_off_trend
+0.2
lookahead_spot
+1.9
travel_km
0.0
baseline
−0.5
LEAN +20.2→ 90% GEORGIA
#7 Miami @ Pittsburgh
Sat Nov 29 · line: Pittsburgh +6.5 · ATS: Miami -6.5
70%
MIAMI ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+8.4
rest_diff
+0.0
opp_pass_def_trend
+0.1
rush_def_trend
+0.2
pass_off_trend
+1.3
lookahead_spot
+1.7
travel_km
−1.4
baseline
−0.5
LEAN +7.8→ 70% MIAMI
#8 Notre Dame @ Stanford ATS SPLIT
Sun Nov 30 · line: Stanford +32.5 · ATS: Stanford +32.5
96%
NOTRE DAME ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+36.7
rest_diff
+0.0
opp_pass_def_trend
−0.2
rush_def_trend
−0.4
pass_off_trend
−1.3
lookahead_spot
0.0
travel_km
−2.7
baseline
−0.5
LEAN +29.6→ 96% NOTRE DAME
⚠ ATS side is STANFORD +32.5. Straight-up the robot takes Notre Dame; against the number, the value is on the other side. It's math, not a mood.
#9 LSU @ Oklahoma ATS SPLIT
Sat Nov 29 · line: Oklahoma -11.5 · ATS: LSU +11.5
75%
OKLAHOMA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+5.2
rest_diff
+0.0
opp_pass_def_trend
+0.2
rush_def_trend
−0.0
pass_off_trend
−0.1
lookahead_spot
+1.7
travel_km
+0.7
baseline
+0.5
LEAN +10.2→ 75% OKLAHOMA
⚠ ATS side is LSU +11.5. Straight-up the robot takes Oklahoma; against the number, the value is on the other side. It's math, not a mood.
#10 Iowa @ Nebraska
Fri Nov 28 · line: Nebraska +6.0 · ATS: Iowa -6.0
71%
IOWA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+11.3
rest_diff
+0.0
opp_pass_def_trend
−0.4
rush_def_trend
+0.0
pass_off_trend
−0.1
lookahead_spot
+0.3
travel_km
−0.4
baseline
−0.5
LEAN +8.2→ 71% IOWA
#11 Utah @ Kansas
Fri Nov 28 · line: Kansas +10.0 · ATS: Utah -10.0
83%
UTAH ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+18.3
rest_diff
−0.0
opp_pass_def_trend
−0.0
rush_def_trend
−0.5
pass_off_trend
+0.3
lookahead_spot
+0.3
travel_km
−1.3
baseline
−0.5
LEAN +14.5→ 83% UTAH
#12 Clemson @ South Carolina
Sat Nov 29 · line: South Carolina -2.5 · ATS: South Carolina -2.5
61%
SOUTH CAROLI ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+0.2
rest_diff
+0.0
opp_pass_def_trend
+0.2
rush_def_trend
+0.5
pass_off_trend
+1.1
lookahead_spot
−0.6
travel_km
+0.2
baseline
+0.5
LEAN +4.0→ 61% SOUTH CAROLINA
#13 Ole Miss @ Mississippi State
Fri Nov 28 · line: Mississippi State +7.0 · ATS: Ole Miss -7.0
87%
OLE MISS ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+18.1
rest_diff
+0.0
opp_pass_def_trend
−0.3
rush_def_trend
+0.8
pass_off_trend
+0.6
lookahead_spot
+1.6
travel_km
−0.1
baseline
−0.5
LEAN +18.2→ 87% OLE MISS
#14 Penn State @ Rutgers
Sat Nov 29 · line: Rutgers +14.0 · ATS: Penn State -14.0
90%
PENN STATE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+20.5
rest_diff
−0.0
opp_pass_def_trend
+0.2
rush_def_trend
+0.4
pass_off_trend
+1.8
lookahead_spot
+0.8
travel_km
−0.3
baseline
−0.5
LEAN +20.9→ 90% PENN STATE
#15 Indiana @ Purdue
Sat Nov 29 · line: Purdue +28.5 · ATS: Indiana -28.5
97%
INDIANA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+33.5
rest_diff
+0.0
opp_pass_def_trend
−0.1
rush_def_trend
+0.1
pass_off_trend
+0.7
lookahead_spot
+2.4
travel_km
−0.1
baseline
−0.5
LEAN +34.1→ 97% INDIANA
#16 Texas Tech @ West Virginia ATS SPLIT
Sat Nov 29 · line: West Virginia +24.0 · ATS: West Virginia +24.0
90%
TEXAS TECH ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+24.8
rest_diff
−0.0
opp_pass_def_trend
−0.0
rush_def_trend
−0.2
pass_off_trend
−1.0
lookahead_spot
+1.2
travel_km
−1.8
baseline
−0.5
LEAN +20.4→ 90% TEXAS TECH
⚠ ATS side is WEST VIRGINIA +24.0. Straight-up the robot takes Texas Tech; against the number, the value is on the other side. It's math, not a mood.
#17 Missouri @ Arkansas ATS SPLIT
Sat Nov 29 · line: Arkansas +4.5 · ATS: Arkansas +4.5
59%
MISSOURI ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+6.2
rest_diff
+0.0
opp_pass_def_trend
−0.1
rush_def_trend
−0.1
pass_off_trend
+0.3
lookahead_spot
−0.1
travel_km
−0.3
baseline
−0.5
LEAN +3.4→ 59% MISSOURI
⚠ ATS side is ARKANSAS +4.5. Straight-up the robot takes Missouri; against the number, the value is on the other side. It's math, not a mood.
#18 Arizona @ Arizona State
Sat Nov 29 · line: Arizona State +2.0 · ATS: Arizona State +2.0
58%
ARIZONA STAT ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+0.1
rest_diff
−0.0
opp_pass_def_trend
+0.1
rush_def_trend
+0.4
pass_off_trend
+0.3
lookahead_spot
−0.6
travel_km
+0.1
baseline
+0.5
LEAN +2.9→ 58% ARIZONA STATE
#19 UCF @ BYU
Sat Nov 29 · line: BYU -17.5 · ATS: BYU -17.5
88%
BYU ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+12.3
rest_diff
−0.0
opp_pass_def_trend
+0.1
rush_def_trend
+0.1
pass_off_trend
−0.6
lookahead_spot
+1.5
travel_km
+2.7
baseline
+0.5
LEAN +18.5→ 88% BYU
#20 UCLA @ USC
Sun Nov 30 · line: USC -21.0 · ATS: USC -21.0
94%
USC ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+22.6
rest_diff
+0.0
opp_pass_def_trend
+0.1
rush_def_trend
+0.1
pass_off_trend
−0.3
lookahead_spot
+0.8
travel_km
+0.0
baseline
+0.5
LEAN +25.9→ 94% USC
#21 SMU @ California
Sun Nov 30 · line: California +13.5 · ATS: SMU -13.5
86%
SMU ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+19.7
rest_diff
+0.0
opp_pass_def_trend
−0.0
rush_def_trend
−0.1
pass_off_trend
+0.4
lookahead_spot
+1.9
travel_km
−2.1
baseline
−0.5
LEAN +17.3→ 86% SMU
#22 Kentucky @ Louisville
Sat Nov 29 · line: Louisville -1.0 · ATS: Louisville -1.0
61%
LOUISVILLE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+2.1
rest_diff
+0.0
opp_pass_def_trend
+0.0
rush_def_trend
−0.2
pass_off_trend
−0.4
lookahead_spot
+0.2
travel_km
+0.1
baseline
+0.5
LEAN +4.2→ 61% LOUISVILLE
#23 Cincinnati @ TCU
Sat Nov 29 · line: TCU -3.0 · ATS: TCU -3.0
67%
TCU ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+1.3
rest_diff
+0.0
opp_pass_def_trend
−0.0
rush_def_trend
−0.1
pass_off_trend
+0.7
lookahead_spot
+1.3
travel_km
+1.2
baseline
+0.5
LEAN +6.8→ 67% TCU
#24 Florida State @ Florida
Sat Nov 29 · line: Florida +1.0 · ATS: Florida +1.0
54%
FLORIDA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
−2.4
rest_diff
−0.1
opp_pass_def_trend
+0.2
rush_def_trend
+0.1
pass_off_trend
+1.1
lookahead_spot
0.0
travel_km
+0.2
baseline
+0.5
LEAN +1.5→ 54% FLORIDA
#25 Navy @ Memphis
Fri Nov 28 · line: Memphis -3.5 · ATS: Memphis -3.5
79%
MEMPHIS ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+6.8
rest_diff
−0.0
opp_pass_def_trend
+0.1
rush_def_trend
+0.5
pass_off_trend
+0.9
lookahead_spot
+0.3
travel_km
+1.1
baseline
+0.5
LEAN +12.2→ 79% MEMPHIS
#26 Wake Forest @ Duke ATS SPLIT
Sat Nov 29 · line: Duke -2.5 · ATS: Wake Forest +2.5
55%
DUKE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+1.2
rest_diff
−0.0
opp_pass_def_trend
+0.0
rush_def_trend
−0.4
pass_off_trend
−1.7
lookahead_spot
+0.3
travel_km
+0.1
baseline
+0.5
LEAN +2.0→ 55% DUKE
⚠ ATS side is WAKE FOREST +2.5. Straight-up the robot takes Duke; against the number, the value is on the other side. It's math, not a mood.
#27 Northwestern @ Illinois ATS SPLIT
Sun Nov 30 · line: Illinois -7.0 · ATS: Northwestern +7.0
61%
ILLINOIS ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+1.7
rest_diff
−0.0
opp_pass_def_trend
−0.3
rush_def_trend
−0.1
pass_off_trend
−1.2
lookahead_spot
+1.7
travel_km
0.0
baseline
+0.5
LEAN +4.3→ 61% ILLINOIS
⚠ ATS side is NORTHWESTERN +7.0. Straight-up the robot takes Illinois; against the number, the value is on the other side. It's math, not a mood.
#28 James Madison @ Coastal Carolina ATS SPLIT
Sat Nov 29 · line: Coastal Carolina +24.0 · ATS: Coastal Carolina +24.0
93%
JAMES MADISO ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+25.0
rest_diff
+0.0
opp_pass_def_trend
−0.5
rush_def_trend
+0.6
pass_off_trend
−0.1
lookahead_spot
+1.7
travel_km
−0.5
baseline
−0.5
LEAN +23.8→ 93% JAMES MADISON
⚠ ATS side is COASTAL CAROLINA +24.0. Straight-up the robot takes James Madison; against the number, the value is on the other side. It's math, not a mood.
#29 Army @ UTSA
Sat Nov 29 · line: UTSA -8.5 · ATS: UTSA -8.5
71%
UTSA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+3.1
rest_diff
−0.0
opp_pass_def_trend
−0.2
rush_def_trend
+0.5
pass_off_trend
+0.6
lookahead_spot
−0.2
travel_km
+2.3
baseline
+0.5
LEAN +8.6→ 71% UTSA
#30 Boise State @ Utah State
Fri Nov 28 · line: Utah State +2.0 · ATS: Boise State -2.0
63%
BOISE STATE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+6.4
rest_diff
+0.0
opp_pass_def_trend
−0.1
rush_def_trend
+0.2
pass_off_trend
+0.4
lookahead_spot
+0.8
travel_km
−0.4
baseline
−0.5
LEAN +4.9→ 63% BOISE STATE
#31 San Diego State @ New Mexico
Fri Nov 28 · line: New Mexico +1.0 · ATS: New Mexico +1.0
57%
NEW MEXICO ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
−1.5
rest_diff
+0.0
opp_pass_def_trend
−0.2
rush_def_trend
+0.6
pass_off_trend
+1.2
lookahead_spot
−0.8
travel_km
+0.9
baseline
+0.5
LEAN +2.7→ 57% NEW MEXICO
#32 Colorado @ Kansas State ATS SPLIT
Sat Nov 29 · line: Kansas State -16.5 · ATS: Colorado +16.5
84%
KANSAS STATE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+13.2
rest_diff
+0.0
opp_pass_def_trend
−0.0
rush_def_trend
−0.2
pass_off_trend
−0.4
lookahead_spot
0.0
travel_km
+0.7
baseline
+0.5
LEAN +15.7→ 84% KANSAS STATE
⚠ ATS side is COLORADO +16.5. Straight-up the robot takes Kansas State; against the number, the value is on the other side. It's math, not a mood.
#33 Virginia Tech @ Virginia
Sun Nov 30 · line: Virginia -8.0 · ATS: Virginia -8.0
75%
VIRGINIA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+7.5
rest_diff
+0.6
opp_pass_def_trend
−0.3
rush_def_trend
+0.1
pass_off_trend
−0.6
lookahead_spot
+0.5
travel_km
+0.2
baseline
+0.5
LEAN +10.4→ 75% VIRGINIA
#34 Wisconsin @ Minnesota
Sat Nov 29 · line: Minnesota +2.5 · ATS: Minnesota +2.5
52%
MINNESOTA ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
−1.2
rest_diff
+0.0
opp_pass_def_trend
+0.0
rush_def_trend
−0.7
pass_off_trend
−0.3
lookahead_spot
−0.1
travel_km
+0.3
baseline
+0.5
LEAN +0.6→ 52% MINNESOTA
#35 Houston @ Baylor
Sat Nov 29 · line: Baylor -2.5 · ATS: Baylor -2.5
60%
BAYLOR ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+1.8
rest_diff
+0.0
opp_pass_def_trend
+0.0
rush_def_trend
−0.3
pass_off_trend
+0.1
lookahead_spot
−0.5
travel_km
+0.2
baseline
+0.5
LEAN +3.8→ 60% BAYLOR
#36 Temple @ North Texas
Fri Nov 28 · line: North Texas -20.0 · ATS: North Texas -20.0
95%
NORTH TEXAS ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+19.6
rest_diff
−0.0
opp_pass_def_trend
−0.2
rush_def_trend
−0.4
pass_off_trend
+2.7
lookahead_spot
+0.9
travel_km
+1.9
baseline
+0.5
LEAN +26.9→ 95% NORTH TEXAS
#37 UNLV @ Nevada
Sun Nov 30 · line: Nevada +7.5 · ATS: UNLV -7.5
78%
UNLV ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+13.9
rest_diff
+0.1
opp_pass_def_trend
+0.2
rush_def_trend
+0.3
pass_off_trend
−0.5
lookahead_spot
+0.9
travel_km
−0.5
baseline
−0.5
LEAN +11.9→ 78% UNLV
#38 Toledo @ Central Michigan
Sat Nov 29 · line: Central Michigan +11.5 · ATS: Toledo -11.5
78%
TOLEDO ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
−2.0
rating_gap
+13.6
rest_diff
−0.2
opp_pass_def_trend
−0.0
rush_def_trend
−0.2
pass_off_trend
+0.3
lookahead_spot
+1.3
travel_km
−0.2
baseline
−0.5
LEAN +12.0→ 78% TOLEDO
#39 South Alabama @ Texas State ATS SPLIT
Sat Nov 29 · line: Texas State -10.0 · ATS: South Alabama +10.0
71%
TEXAS STATE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+6.1
rest_diff
−0.0
opp_pass_def_trend
−0.0
rush_def_trend
−0.4
pass_off_trend
−0.0
lookahead_spot
−0.6
travel_km
+0.8
baseline
+0.5
LEAN +8.5→ 71% TEXAS STATE
⚠ ATS side is SOUTH ALABAMA +10.0. Straight-up the robot takes Texas State; against the number, the value is on the other side. It's math, not a mood.
#40 Western Kentucky @ Jacksonville State
Sat Nov 29 · line: Jacksonville State +1.5 · ATS: Jacksonville State +1.5
71%
JACKSONVILLE ▾
THE RECEIPT — SIGNALS SUM TO THE LEAN, THE LEAN IS THE PROBABILITY
home_field
+2.0
rating_gap
+3.9
rest_diff
+0.0
opp_pass_def_trend
−0.1
rush_def_trend
−0.2
pass_off_trend
+1.5
lookahead_spot
+0.5
travel_km
+0.3
baseline
+0.5
LEAN +8.3→ 71% JACKSONVILLE STATE
⛏ THE SIGNAL MINE
15 CANDIDATES · EVERYONE AUDITIONS · FEW ARE CAST
A clue goes live only if it improves held-out predictions across 3+ seasons with a stable coefficient sign. Cut signals stay in the repo with their evidence — the robot keeps an audition record on everyone, including itself.
✓ GRADUATED — LIVE IN THE MODEL
home_field rating_gap rest_diff opp_pass_def_trend rush_def_trend pass_off_trend lookahead_spot travel_km
⏳ AUDITIONING — 3-SEASON TRIAL IN PROGRESS
talent_gap 2/3 returning_production 2/3 line_movement 0/3
✗ CUT — EVIDENCE ON FILE
rush_off_trend — no lift on held-out seasons (mean Brier delta -0.00003)
turnover_luck — no lift on held-out seasons (mean Brier delta -0.00002)
letdown_spot — no lift on held-out seasons (mean Brier delta -0.00014)
body_clock — no lift on held-out seasons (mean Brier delta -0.00014)
📋 SCOREBOARD INCL. THE LOSSES
2025 season vs closing line (ATS)336–335–14 · 50.1%
straight-up496–189 (rigged, ignore)
career Brier (2021+).190 baseline .191
career vs closing line1,954–1,915–75 (50.5%)
worst weekWeek 10 · 17–35
🎯 CALIBRATION
40–50%
47.1% n=444 50–60%
52.6% n=521 59–69%
65.0% n=549 70–80%
72.4% n=557 80–90%
85.6% n=536 90–100%
93.0% n=359
Career held-out calibration, 2021+. When I say 70%, check the bar — that's the arrangement.
SIGNAL LEADERBOARD · WHEN DOMINANT CLUE
travel_km9–8 · 52.9%
rating_gap287–283–11 · 50.4%
home_field37–38–3 · 49.3%
pass_off_trend2–3 · 40.0%
🪦 WHAT I GOT WRONG LAST WEEK

I gave Army a 93% win probability against Tulsa. Army lost by 1. The dominant clue was rating_gap (+1.8 of lean), which has been overruled by reality. The receipt stays up; that is the arrangement.

I gave Marshall a 71% win probability against App State. Marshall lost by 2. The dominant clue was rating_gap (-1.0 of lean), which has been overruled by reality. The receipt stays up; that is the arrangement.

GENERATED FROM THE SIGNAL CONTRIBUTIONS · NO HUMANS WERE CONSULTED · FULL SEASON OF CONFESSIONS ▸
📜 THE HISTORY — BACKTEST PICKS VS WHAT ACTUALLY HAPPENED
Every season below was predicted walk-forward: the model only ever saw seasons before the one it was picking. These are the receipts for the whole operation.
SEASONGAMESSTRAIGHT-UPVS THE LINEATS %BRIERML ROI (FLAT $)
2016 76070.3%358–382–2048.4%.190 no ML data
2017 77670.4%390–365–2151.7%.185 no ML data
2018 77272.4%391–361–2052.0%.175 no ML data
2019 77473.1%377–384–1349.5%.176 no ML data
2020 (COVID — reported, not counted) 53471.2%274–251–952.2%.187 no ML data
2021 77069.5%383–378–950.3%.187 -4.0%
2022 77668.7%389–374–1351.0%.199 -4.6%
2023 79272.3%395–381–1650.9%.185 -1.2%
2024 79869.2%391–388–1950.2%.193 -5.4%
2025 80872.5%396–394–1850.1%.184 +1.0%
💸 The moneyline question, answered on 3,771 games (2021-2025): betting every straight-up pick flat returns -2.8% per dollar — the vig eats it. The one live thread: when the robot backs an UNDERDOG straight up (470 bets), flat ROI is +11.4%, and a ≥5-point modeled edge returns +2.9% on 968 bets. That's the size of sample where luck still writes headlines — so it's a paper-ledger experiment, not a bankroll strategy. The ledger will decide.
🧪 THE LAB — BUILD A WORSE ROBOT
10 SEASONS RE-RUN ON EVERY TOGGLE
HOME FIELD ✓ RATING GAP ✓ REST DIFF ✓ OPP PASS DEF TREND ✓ RUSH DEF TREND ✓ PASS OFF TREND ✓ LOOKAHEAD SPOT ✓ TRAVEL KM ✓
—
ATS WIN RATE
—
BRIER SCORE
—
10-SEASON ATS
ROBOT POLL · rebuilt every Tuesday by a cron job on a rented computer · picks lock at publish and are never edited
2020 reported separately (you know why) · the robot bets fake money so you can lose real money responsibly