The Sports PageMississippi State Won by 25. The Numbers Say Minnesota Is Just Bad.
Mississippi State beat Minnesota 38–13. The ratings moved Mississippi State up 0.6 and moved Minnesota down 4.9. Same game. Almost all of it was about the team that lost.
What happened?
Mississippi State went to Minnesota and won 38–13. Nobody saw it coming. The betting line had the game dead even.
Most people read that one way: Mississippi State must be good.
Here is the surprising part. You would expect a 25-point win to say most about the team that won it. It did not. The ratings moved the loser eight times further than they moved the winner — and the betting line for this week read the game the other way round.
Is that right?
Maybe not. There is a second reading, and the ratings prefer it. Minnesota might just be bad.
Every result has two teams in it. A blowout can mean the winner is strong or the loser is weak, and the scoreboard does not tell you which.
So which was it?
The rating barely moved the winner. It dropped the loser by almost five points.
Read that again, because it is the whole piece. One result, and the number that is supposed to measure teams decided it was mostly about Minnesota.
Every game has two teams in it. Ask which one the result was about before you decide what it proved.
The ProfessorWhy would the rating think that?
Because of what Minnesota had done before. Look at their whole season:
| Minnesota, 2026 | Result |
|---|---|
| Eastern Illinois | Won 59–7 |
| Mississippi State | Lost 13–38 |
That is it. Two games.
The first was against a team from the division below. Nobody learns anything from those; we do not even rate them. So before Saturday, Minnesota had played no games that told you anything.
Which means the even betting line was not built on what Minnesota had done. It was built on what people expected in July. Then a real game happened, and it was bad.
Did Mississippi State just buy a team?
We checked, because it is a fair thing to wonder now that players move so freely.
| Players in | How good were the teams they left | Players back from last year | |
|---|---|---|---|
| Mississippi State | 28 | +7.5 | 34% |
| South Carolina | 27 | +6.9 | 69% |
Almost the same. Both took a whole new class of players from teams of about the same quality. Mississippi State is not hiding anything South Carolina is not doing too.
One thing does stand out. Mississippi State kept only 34% of last year’s production; South Carolina kept 69%. That cuts both ways. A new team can be better than its old record says — and after two games, you still barely know it.
Does a blowout mean anything for next week?
We checked three seasons. Teams that beat the number by 20 or more went 47.9% against the spread the week after.
That is a shade under half, which hints the market gets a little carried away. But the honest range runs from 44% to 52%, so we cannot say it for certain. Call it a lean, not a fact.
What to take home
When a team wins big, ask one question before you believe it: who was that game about?
Look at what the loser had done before. If they had not really been tested, the blowout may be telling you about them, not about the winner.
Nobody asks this on Saturday night. It is the cheapest question in sports and almost no one uses it.
Baseball is ranked by championship leverage: we simulate the rest of the season 60,000 times, then split those seasons by who won each of today's games. College football cannot be simulated that way, so it is ranked by how close to a coin flip the ratings make it — a lopsided game teaches you nothing. Tomorrow this block reports what happened to today's pick.
Notes & sources
SP+ ratings before the game were frozen on 2026-09-10 by scripts/snapshot_ratings.py, so the "before" column cannot have seen the result. Mississippi State +9.4 to +10.0, Minnesota +9.6 to +4.7, South Carolina +16.7 to +16.9.
Lines are the mean across books via the CollegeFootballData API. Minnesota–Mississippi State closed at Minnesota −0.33, effectively a pick. Week 3 has South Carolina −4.0 at home against Mississippi State.
Transfers from the CFBD portal feed, 2026: 28 into Mississippi State, 27 into South Carolina; the figure quoted is the mean SP+ of the programmes those players left. Returning production is the share of last season’s output back on the roster.
The blowout follow-up covers 2023–2025, 6,626 consecutive team-week pairs with a line in both weeks. Teams beating the closing number by 20+ went 307–334 the following week, 47.9%, with a 95% interval of 44.0% to 51.8%. Correlation between a team’s cover margin one week and the next: r = −0.006.
Three limits. SP+ is one rating system and others weight recency differently, so "the ratings" here means this one. Two games is a tiny sample for any of these clubs, Mississippi State included — the argument is about what a result can support, not about who wins on Saturday. And the 47.9% lean is not separable from 50%; if it were a finding rather than a hint, we would have led with it.
