Does Beating the Betting Line Once Mean a Team Will Do It Again?
PvE Top 25 next-game results, Weeks 2 to 8
Ranktory's PvE Top 25 are the teams that have beaten the oddsmakers' expectations by the most. We checked how they did in their next game, and gave the PvE rankings a new look to show more of the story.
· Ranktory Research · Polls: AP and Coaches, as published each week, Weeks 2 to 8 · Data: game results through week 8
Before every college football game, oddsmakers publish a point spread: their best guess at the final margin. If a team is favored by 10 and wins by 24, it did 14 points better than expected. If it wins by only 3, it did 7 points worse, even though it won.
Performance vs. Expectation, or PvE, is that gap, averaged across a team's games. A positive number means a team has been beating what the market expected. A negative number means it has been falling short. The PvE rankings put all 138 FBS teams in order by it.
That raises an obvious question. If a team keeps beating expectations, should we expect it to keep doing it?
We tested it. Before each week from 2 through 8, we took the PvE Top 25 and tracked how those teams did against the spread in their very next game. Going in, they had averaged +13.8 points per game better than expected. In that next game, they averaged +0.59.
Across all 161 of those games, they covered the spread 81 times, missed 79 times, and pushed once. That is a 50.6% cover rate, which is what a coin flip produces. (A "cover" is beating the spread. A "push" is landing exactly on it.)
So what does a run of beating the line really tell us? To look at that properly, we rebuilt how the PvE rankings are displayed.
Think of it as an earnings report
If you follow the stock market, there is a familiar way to think about it.
Before every game, the betting line is the price: the market's estimate of the final margin. The final score is the earnings report. PvE is the earnings surprise, the gap between what the market expected and what actually happened.
That comparison is useful. It also comes with a familiar warning. A company that beats estimates this quarter isn't guaranteed to beat them next quarter, because the people setting the estimates see the results too. Whether the same thing happens to a football line is exactly the question we want to test, and we'll come back to it.
The new PvE rankings
Until now, each team's row gave you an average, a Swing and a Trend. The PvE rankings now read more like a scorecard, with the columns below.
| Team | Rank | Average PvE | Vs the line | Median game | Avg line | Form | Trend | Swing |
|---|---|---|---|---|---|---|---|---|
| Oregon State | 1 | +16.32 | 6‑1‑0 | +14.50 | +3.2 | Falling | 12.0 | |
| Wake Forest | 2 | +14.18 | 5‑2‑0 | +14.25 | +3.8 | Steady | 15.4 | |
| New Mexico | 3 | +13.39 | 6‑1‑0 | +13.25 | +12.3 | Falling | 11.9 | |
| UCLA | 3 | +13.39 | 7‑0‑0 | +8.00 | +5.8 | Falling | 16.7 | |
| Northwestern | 5 | +10.82 | 5‑2‑0 | +12.50 | +7.3 | Falling | 13.4 |
Beat the line Missed One bar per game, oldest on the left. Bar height shows by how much.
Rotate your phone sideways for the full view.
- Average PvE. The average surprise per game, which still sets the rank.
- Vs the line. A beat-miss-push record against the line. The line is the average of the spreads from the sportsbooks that posted one (for now, DraftKings and Bovada as reported by CFBD). It is not any single book's number, and it is not a verified closing line. A game counts only when at least two books posted a spread. A beat means PvE above zero, a miss below zero, and a push exactly zero.
- Median game. The typical game, so one blowout can't carry the whole row.
- Avg line. How many points the team has been favored (+) or an underdog (−) by, on average. It tells you how much the market expected going in.
- Form. One bar for each game, oldest on the left. Green beat the line, red missed, and the height shows by how much.
- Trend and Swing. The same measures as before. Trend asks whether the last four games are rising or falling by more than a point per game. Swing shows how far single games stray from the team's average.
Where the average and the record agree
UCLA has beaten the line in all seven of its counted games and averages +13.39, tied for third in the country. Oregon State, first at +16.32, is 6-1, and its form bars show a +37.50 game followed by a −0.50 that makes its Trend read Falling. A record of 6-1 and a big average point the same way.
Where they don't
An average can look the same for very different teams. The extra columns show why.
Arizona averages +5.96, 17th in the country. The Wildcats are 6-1 on the field, but their record against the line is 3-4, and their median game is −1.75. Its three most recent games, +20.25, +17.50 and +24.00, lift the average, but in the typical game Arizona has finished short of the market.
LSU averages +6.22, 15th, and is 4-4. Its average line is +17.9, so the market has generally expected the Tigers to win comfortably, and they have beaten those expectations as often as they've missed them. Alabama is also 4-4 at +4.88. Its last two games came in at −23.75 and −2.25, which is where its Falling Trend comes from.
It works the other way around too. Washington State ranks 103rd at −2.59, yet it has beaten the line in five of eight games. And Indiana, first in our Performance Index Poll, is 1-7 against the line and averages −3.44, 108th.
None of that makes Indiana or LSU a bad team. It says the market expected more of them than they delivered in those games, and the record and the bars show how often, not just by how much.
Does a good line record carry over?
Back to the opening number. Before Week 9, we tested it using only information a reader would have had at the time. For each week from 2 through 8, we took the PvE Top 25 as published after the previous week and tracked how those teams did against the line in their next game.
| Group | Covers | Misses | Pushes | Cover rate | 95% interval |
|---|---|---|---|---|---|
| PvE Top 25 | 81 | 79 | 1 | 50.6% | 43.0% to 58.3% |
| AP Top 25 | 77 | 77 | 1 | 50.0% | 42.2% to 57.8% |
| Coaches Top 25 | 73 | 82 | 1 | 47.1% | 39.4% to 54.9% |
| All teams outside the PvE Top 25 | 343 | 336 | n/a | 50.5% | 46.8% to 54.3% |
A cover is beating the line, a miss is falling short of it, and a push is landing exactly on it. Cover rate leaves pushes out.
The PvE Top 25 did not do better than the polls, and it did not do worse. Splitting it up doesn't change the picture. Ranks 1 to 10 covered 33 of 63 (52.4%). Ranks 11 to 25 covered 48 of 97 (49.5%). Against opponents outside the Top 25, the group went 72-70.
The weekly results swing more than you might expect.
| Games played in | Covers | Misses | Cover rate |
|---|---|---|---|
| Week 2 | 12 | 13 | 48.0% |
| Week 3 | 8 | 17 | 32.0% |
| Week 4 | 19 | 8 | 70.4% |
| Week 5 | 10 | 9 | 52.6% |
| Week 6 | 12 | 9 | 57.1% |
| Week 7 | 6 | 16 | 27.3% |
| Week 8 | 14 | 7 | 66.7% |
Weeks 4 and 7 look dramatic. They involve 27 and 22 games, and neither is statistically distinguishable from chance. When you check seven weeks, one or two near-misses are normal.
Put plainly, we found no edge. The scorecard describes what has already happened. It is not a forecast, and it is not betting advice.
What a fan should object to
We would rather you hear the objections from us.
It's a small sample. Yes. That's 160 games from 54 different teams over seven weeks. The interval around 50.6% runs from about 43% to 58%. A real but modest edge, say 55%, would be invisible at this size. We can't say PvE has no predictive value. We can say we don't see any yet.
The line already prices in quality. This is the earnings-report warning again, and it may be the most natural explanation. If a team keeps beating the line, the market adjusts, and by the next game the price already reflects what we're looking at. The data can't separate that from "the ranking simply has no signal," so we won't pick one. Testing how quickly the market catches up is the next study in this series.
Regression to the mean. The drop from +13.8 to +0.59 looks like regression. But the next-game average landed near zero, not below it, so we don't see a reverse effect either. Early-season averages also rest on very few games (the site marks teams with fewer than four).
The line is an average of two books. Correct, so a team's record can differ from what a single sportsbook would show. Pushes are rare under this definition: 10 of 857 team-games in Weeks 2 through 8.
Teams repeat. Fifty-four teams supply the 160 rows. Resampling by team gives an interval of 42.3% to 58.3%, nearly the same as the plain one.
We ran many splits. About twenty. None came close to a significant result, so treat each as description, not discovery.
What we know, and what we don't
What the evidence shows: In 160 next games, the PvE Top 25 covered at about the rate you'd expect from a coin flip. So did the AP and Coaches Top 25 and every PvE rank group we checked.
What it suggests: A high PvE average, by itself, hasn't predicted a cover in the next game this season. The simplest reading is that the market adjusts. The data doesn't prove it.
What remains unresolved: Whether more games reveal a small edge, whether it appears in earlier seasons, and whether the record and the median tell us something the average doesn't. We'll find out as more weeks arrive and as 2024 and 2025 are added.
PvE records how teams performed relative to what the market expected. It was never meant to forecast the next game, and this first test is a useful reminder of that.
Explore the PvE Research Series
Meet PvE: The Teams Exceeding Expectations in 2026
Does Beating the Betting Line Once Mean a Team Will Do It Again? (you are here)
Up next: When does the market catch up?
Explore the Data
Open the new PvE rankings, search any team, and read its record, median game and form next to its average.