The Six-Win Race at Midseason: Record, Margin and What's Left
Minnesota is 4-2 and Penn State 3-3, yet Penn State is likelier to reach six wins. Margins plus the games left beat the record table by 19% in 2024 and 2025.
Section
College Football
The expanded playoff, returning production, and realignment — read through power ratings and play-by-play instead of vibes.
College football measured rather than argued about. The through-line is that most of what gets asserted in a broadcast booth is a testable claim: that a twelve-team playoff changed who contends, that returning production predicts a jump, that a week-one loss is fatal, that turnover margin is a skill. Each piece states its sample, its filters and its formula, replays the games, and reports the answer including the times the answer is “not much.” Where a season is still running, the article carries the date its data was pulled. If you want the machinery behind the ratings used here, start in Stat Explainers; if you want to build it yourself, the Tutorials section has the code.
36 articles, newest first
-
-
What Moves the AP Poll: A Ranked Win, a Close Call or a Loss
Two ranked teams both win: the lower one finishes ahead 18.4% of the time if it alone beat a ranked team, 2.3% if neither did and the upper won by more than a score.
-
The Unbeaten Curve, Three Weeks In: 2026 Is Running Faster
34 of 136 FBS teams are unbeaten through week three, against 2024's 43 of 134. The gap opened in ten days, with six first losses to non-FBS teams.
-
What a College Football Crowd Number Actually Measures
Seven of the 39 hosts that played twice in weeks one and two announced the identical attendance figure both times. It measures the stadium, not the game.
-
What the Line's Week-One Lesson Was Worth
The moved week-two lines finished 0.37 points a game closer than the openers, SE 0.30. The market's biggest reaction was its biggest miss.
-
The Line Learned From Week One, at a Tenth of the Volume
79 Saturday lines from open to game morning: each point a team beat its week-one number by moved its line a tenth of a point, and the moving was done by Wednesday.
-
Ninety-One Games, Thirty-Five Bits: What Week One Told Us
Week one's 91 games were priced to carry 34.8 bits of information and delivered 31.2. Week two's 58 priced games are worth more before a snap.
-
Opening Saturday 2026, Priced: 68 Games Against the Number
Saturday September 5, 2026, measured against the line: favorites 63-5 and 45-23 against the spread, FBS road underdogs 0-26, and a miss that was 16 points wide.
-
The Schedule Already Knew: 2026's Opening Weeknights, Measured
The 19 completed games of 2026's opening weeknights: a rule using only the schedule went 16-3, and all three misses came from the six games with no ranking.
-
What a Week-One Loss Actually Costs: One Game of Arithmetic, and the Rest Is Identity
2024's 37 FBS opener losers, tracked: 46% still reached six wins, close losses predicted no better than blowouts, and the real cost was one game of slack.
-
Week Zero 2026, Measured: Eight Games and No One Left to Ambush
Week zero 2026 was eight FBS-vs-FBS games and zero FCS visitors - so exactly 8 of 16 teams had to lose. The opening weekend's only wildcard is gone, measured.
-
The Week-One Slate Itself: What 2026's Opening Week Was Built to Do
The real 2026 week-one slate, pulled before kickoff: 91 games, a halved Thursday, a tripled Sunday, 48 FCS visitors, and exactly one ranked-vs-ranked game.
-
The Franklin Reset: Virginia Tech’s 2026 Season, Measured
Franklin took over 2-10 Vanderbilt and a sanctioned Penn State; now a 3-9 VT outscored by 8.8 a game, facing a 2026 slate ramping from .347 to .673 opponents.
-
The Last Unbeaten: How 134 Perfect Records Become One
Replaying 877 games of 2024: 134 unbeatens become 96 in a week and 1 by November 23 - and 89 of 133 first losses came from another unbeaten, not an upset.
-
When the Record Starts to Mean Something: A Raw Record Loses to a Coin Flip Until November
A raw record predicts the rest of a season worse than a .500 guess until game 8; shrunk, it helps from game 3. Plus when results outweigh the preseason poll.
-
Opening Week, Measured: 100 Games in Ten Days, and Almost No Signal
2024's opening week, measured: 100 games in ten days, 61 of them FCS buy games, blowouts at 43.6%, and openers that explained just 10% of rest-of-season margin.
-
The FCS Buy Game, Measured: 121 Paid Visits, a 5% Upset Rate
121 FBS-vs-FCS buy games in 2024: 62.8% of opening week, a 31-point median margin, 17 shutouts, and a 5% upset rate that fell entirely on the Group of Five.
-
Weeknight Football, Measured: Tuesday Plays Close, Thursday Plays Ugly
Tuesday MACtion played the closest football of 2024, Thursday ran 60% blowouts, and Friday beat Saturday for drama. The 134 weekday games, measured.
-
The Bye Week, Measured: Testing the Lead We Flagged
A bye alone was worth +1.2 points in 2024 - noise. The blowout-then-bye lead flagged August 5 held at +4.4 (SE 1.3) through every stress test one season allows.
-
The Letdown and the Bounce-Back: Two Myths, One Audit
After 327 blowout wins, 2024 teams played +1.4 points above their own baseline; after 235 blowout losses, -0.1. Neither the letdown nor the bounce-back is real.
-
The Fourth-Quarter Lie: What 877 Games Say About Comebacks
Teams leading after Q3 won 87.1% of 2024's 877 cached games, and 25-point leads went 189 for 189. The real comeback rates by deficit, and who owns garbage time.
-
Who Goes to Overtime? The 35-Game Census of a 4% Lottery
Overtime hit 35 of 877 games in 2024, a flat 4% that barely moves by month, conference, or ranking. The hour before the tie, and the 8-OT box score, decoded.
-
How Lopsided Is College Football? 877 Games, Sorted by Final Margin
Just over a third of 2024's 877 games were one-score and 29% were decided by 25+ points. The full margin distribution, and a team's margin trend in Python.
-
How Often Does the Better Offense Win? A 532-Game Answer
The team with the higher scoring average entering kickoff won 67.3% of 532 cached 2024 matchups — but barely 57% when the gap was under 3 points a game.
-
Home-Field Advantage, Measured: How +8.8 Points Becomes +3.0
The home team won 64% of 2024 FBS games by an average of +8.8 points, but the real edge is about +3. Measured from 873 cached games, schedule bias split out.
-
Yards Per Play Built a Champion: Ohio State's 2024 Season, Game by Game
Ohio State's two 2024 regular-season losses were its two closest games by yards per play — and against Michigan it won the per-play battle and lost anyway.
-
How College Football Overtime Math Works (and Why Going Second Wins)
College overtime gives each team a possession from the 25 — and the team playing defense first wins about 56% of the time. We model that information edge.
-
Recruiting Geography: Where College Football's Talent Comes From
Talent isn't spread evenly. Where the blue-chip recruits actually come from, why a few states dominate, and what that means for who can contend.
-
Special Teams in College Football: The Hidden Points
Field goals, returns, and field position quietly swing games. How to value the third phase, why it's volatile, and where the hidden yardage really adds up.
-
Second-Order Wins: Measuring How Lucky a College Football Team Got
How many games should a team have won? Postgame win expectancy turns the box score into expected wins — and the gap from the real record is mostly luck.
-
Tempo in College Football: Do Faster Offenses Really Win More?
Up-tempo offenses run more plays — but more plays isn't the same as more efficiency. What pace does and doesn't do, measured per play instead of per game.
-
How Recruiting Rankings Work: Star Ratings, Composites, and What They Predict
What star ratings really mean, how the 247Sports Composite and team class rankings are built, and how much recruiting actually predicts on the field.
-
The Blue-Chip Ratio: Does Recruiting Talent Actually Predict Titles?
The blue-chip ratio is the simplest recruiting-to-results stat there is. What it measures, why every recent champion clears the bar, and where it breaks down.
-
The 12-Team Playoff Era, by the Numbers: What's Actually Changed
Two seasons of 12-team College Football Playoff results, scored: seed performance, byes, blowouts, and what the bracket really rewards.
-
Returning Production: How Much Does It Really Predict? A Data Check
Returning production correlates with next-year results — but how strongly, and where does the relationship break down?
-
Conference Realignment's Statistical Aftermath: Travel, Parity, and Power Ratings
The new superconferences, measured: cross-country travel miles, competitive parity, and what realignment did to the power curve.