We simulate every game on the framework that actually explains college football — explosiveness and efficiency drive the projected margin, scoring rates set the total — then post only the handful we have the most conviction in. Every posted game gets a grade, a projected score, and a position-by-position breakdown, and every one is graded against the real result.
2026 season to date
Graded games
91.7%
the games we post · 12
Full slate
75.3%
every game · 97
Margin error
17.4
points, full slate
On the board
6max
per week · B+ or better
Accuracy is measured against the actual final score — did we pick the winner, and how many points off was the margin — not against the betting line. Across 2019-2024 the simulator calls about 69% of all games correctly, and roughly 91% of the graded ones we actually post, holding up in every season including those it wasn’t built on. It is a prediction tool, not a betting edge: it does not beat the spread. Analysis and entertainment, 21+, 1-800-GAMBLER.
Our 6 highest-conviction calls out of 58 games simulated this week. We post at most 6, and nothing graded below a B+ — a thin week posts fewer games rather than padding the board.
Kent State
Ohio State
11
38
Explosiveness
Ohio State
Efficiency
Ohio State
Trenches
Ohio State
+3.2 at the line
Dru Deshields
0.34 PPA/play
Julian Sayin
Exactly what we posted last week, graded against the real final score — wins and misses both.
The grade is our conviction in our own call — how big the projected margin is, and whether explosiveness, efficiency, and the trenches all agree with it. Here is how each grade actually performed across 2019-2024:
Explosiveness and efficiency — measured as opponent-adjusted predicted points added, on both sides of the ball — set the projected margin. Scoring rates set the total. Every rating is computed as of that game’s kickoff, using only games already played, so nothing is graded with hindsight.
The trenches are shown, not baked in. We tested whether O-line / D-line play — line yards, short-yardage push, and stuff rate — improves the projection. On four seasons the model was never fit on, it moved winner accuracy from 69.5% to 69.4% and made the margin error slightly worse. Line play is already inside the efficiency numbers, so adding it again just adds noise. We show who wins at the line because it’s a real part of the story — we just don’t let it move the pick when the data says it shouldn’t.
Weeks 1-3 carry in last season. A team needs three games before its current-year rating means anything, so until then we use its full prior-season rating. Tested on 2019, 2020, 2021, 2024, 2025, the early board came in at 91.5% straight up over 82 games — the winner call survives the roster turnover, the projected score gets noticeably looser, and games marked carry-in are the ones still using it.
The board turns over Tuesday at noon ET. A week’s games stay up all the way through the following Monday, collecting their own grades as the finals land, and the next week’s board replaces them Tuesday midday. Last week’s calls stay on this page after the roll, wins and misses both.
0.62 PPA/play
Mason Blakemore◄
0.22 PPA/play
Bo Jackson
0.13 PPA/play
Da'Realyst Clark
0.75 PPA/play
Jeremiah Smith◄
1.19 PPA/play
Terik Mulder
0.93 PPA/play
Nate Roberts◄
1.66 PPA/play
Bad
3.12 line yds
Average◄
3.12 line yds
Bad
10% stuff rate
Good◄
21% stuff rate
Average
9% front-7 havoc
Good◄
12% front-7 havoc
Bad
1.32 2nd-lvl yds
Average◄
1.08 2nd-lvl yds
Bad
48% pass succ allowed
Good◄
29% pass succ allowed
Good
1.41 expl. allowed
Good◄
1.41 expl. allowed
Grades are percentiles against all FBS teams this season. QB/RB/WR/TE grade the team’s primary player at that spot on his own production; the six marked unit have no public per-player data, so they grade the unit’s on-field production (line yards, stuff rate, havoc, coverage) rather than individual players. Not PFF grades.
Michigan State
Notre Dame
18
39
Explosiveness
Notre Dame
Efficiency
Notre Dame
Trenches
Notre Dame
+1.2 at the line
Alessio Milivojevic
0.34 PPA/play
C.J. Carr◄
0.55 PPA/play
Kenneth Williams
1.02 PPA/play
Nolan James Jr.◄
0.55 PPA/play
Rodney Bullard Jr.
0.48 PPA/play
Jordan Faison◄
0.88 PPA/play
Jayden Savoury◄
0.49 PPA/play
James Flanigan
0.43 PPA/play
Average
3.24 line yds
Good◄
2.77 line yds
Good◄
27% stuff rate
Good
23% stuff rate
Good
15% front-7 havoc
Elite◄
19% front-7 havoc
Bad
1.37 2nd-lvl yds
Elite◄
0.50 2nd-lvl yds
Elite◄
31% pass succ allowed
Average
41% pass succ allowed
Bad
1.73 expl. allowed
Bad◄
1.65 expl. allowed
Grades are percentiles against all FBS teams this season. QB/RB/WR/TE grade the team’s primary player at that spot on his own production; the six marked unit have no public per-player data, so they grade the unit’s on-field production (line yards, stuff rate, havoc, coverage) rather than individual players. Not PFF grades.
Houston
Texas Tech
15
36
Explosiveness
Texas Tech
Efficiency
Texas Tech
Trenches
Texas Tech
+0.7 at the line
Conner Weigman◄
0.60 PPA/play
Will Hammond
0.26 PPA/play
Makhi Hughes
0.27 PPA/play
J'Koby Williams◄
0.37 PPA/play
Amare Thomas◄
0.85 PPA/play
Kenny Johnson
0.83 PPA/play
Patrick Overmyer◄
0.72 PPA/play
Matt Ludwig
0.94 PPA/play
Elite◄
4.16 line yds
Average
3.06 line yds
Elite◄
28% stuff rate
Average
16% stuff rate
Good◄
13% front-7 havoc
Good
9% front-7 havoc
Elite◄
0.16 2nd-lvl yds
Good
0.77 2nd-lvl yds
Elite◄
39% pass succ allowed
Elite
31% pass succ allowed
Average◄
1.54 expl. allowed
Bad
1.93 expl. allowed
Grades are percentiles against all FBS teams this season. QB/RB/WR/TE grade the team’s primary player at that spot on his own production; the six marked unit have no public per-player data, so they grade the unit’s on-field production (line yards, stuff rate, havoc, coverage) rather than individual players. Not PFF grades.
Northern Illinois
Arizona
13
33
Explosiveness
Arizona
Efficiency
Arizona
Trenches
Arizona
+1.2 at the line
Taron Dickens
0.06 PPA/play
Noah Fifita◄
0.37 PPA/play
Elijah Porter
0.12 PPA/play
Kedrick Reescano◄
0.15 PPA/play
DeAree Rogers
0.38 PPA/play
Brandon Phelps◄
3.31 PPA/play
Jason Fowler
0.10 PPA/play
Cole Rusk◄
1.38 PPA/play
Bad
2.38 line yds
Elite◄
3.80 line yds
Average◄
17% stuff rate
Bad
24% stuff rate
Average
6% front-7 havoc
Good◄
12% front-7 havoc
Bad
1.31 2nd-lvl yds
Average◄
1.00 2nd-lvl yds
Average
43% pass succ allowed
Good◄
32% pass succ allowed
Good◄
1.29 expl. allowed
Bad
1.69 expl. allowed
Grades are percentiles against all FBS teams this season. QB/RB/WR/TE grade the team’s primary player at that spot on his own production; the six marked unit have no public per-player data, so they grade the unit’s on-field production (line yards, stuff rate, havoc, coverage) rather than individual players. Not PFF grades.
Syracuse
Pittsburgh
19
38
Explosiveness
Pittsburgh
Efficiency
Pittsburgh
Trenches
Pittsburgh
+0.5 at the line
Steven Angeli
0.20 PPA/play
Mason Heintschel◄
0.44 PPA/play
Tylik Hill
0.65 PPA/play
Justin Cook◄
2.95 PPA/play
Umari Hatcher
0.65 PPA/play
Censere Lee◄
1.16 PPA/play
David Clement◄
2.76 PPA/play
Elijah Lagg
1.88 PPA/play
Good◄
2.94 line yds
Bad
2.64 line yds
Good
27% stuff rate
Elite◄
24% stuff rate
Elite◄
16% front-7 havoc
Good
12% front-7 havoc
Good◄
0.67 2nd-lvl yds
Good
0.76 2nd-lvl yds
Good
28% pass succ allowed
Good◄
34% pass succ allowed
Good◄
1.41 expl. allowed
Bad
1.75 expl. allowed
Grades are percentiles against all FBS teams this season. QB/RB/WR/TE grade the team’s primary player at that spot on his own production; the six marked unit have no public per-player data, so they grade the unit’s on-field production (line yards, stuff rate, havoc, coverage) rather than individual players. Not PFF grades.
North Dakota State
Sacramento State
34
16
Explosiveness
North Dakota State
Efficiency
North Dakota State
Trenches
North Dakota State
+4.0 at the line
Nathan Hayes◄
0.30 PPA/play
Jackson Sharman
0.22 PPA/play
DJ Scott
0.23 PPA/play
Isaiah Ene◄
0.36 PPA/play
Jackson Williams◄
0.88 PPA/play
Onterrio Smith Jr.
0.51 PPA/play
Reis Kessel◄
0.33 PPA/play
Kyle Koontz
0.91 PPA/play
Average◄
2.87 line yds
Bad
2.48 line yds
Average
15% stuff rate
Good◄
20% stuff rate
Good◄
9% front-7 havoc
Average
12% front-7 havoc
Average
0.78 2nd-lvl yds
Good◄
0.79 2nd-lvl yds
Good◄
35% pass succ allowed
Bad
42% pass succ allowed
Good◄
1.50 expl. allowed
Average
1.51 expl. allowed
Grades are percentiles against all FBS teams this season. QB/RB/WR/TE grade the team’s primary player at that spot on his own production; the six marked unit have no public per-player data, so they grade the unit’s on-field production (line yards, stuff rate, havoc, coverage) rather than individual players. Not PFF grades.
Winner accuracy by grade, 4,655 games. We post at most 6 a week and never below a B+, which lands around 90.8% — but that is because a high grade means a clearer mismatch, and the books know it too. Higher accuracy here does not mean a betting edge.
