Only so much can change in the first three-plus weeks of a college football season. Most top teams will play one decent opponent at most in that span. Most of the stuff we will end up remembering from this season hasn't happened yet.
Of 133 FBS teams in 2023, two-thirds of them (89) have a current projected win total (per SP+) within one win of where it was in the preseason. The top three teams in the preseason AP poll (Georgia, Michigan and Ohio State) have outscored their first nine opponents by a combined score of 334-60. Their respective outlooks and expectations have barely changed, if at all. Some of the biggest questions we entered the season with -- Can Ohio State's defense rein in big plays against the best opponents? Will USC's defense show up? Is Drew Allar ready to make Penn State a contender? -- have partial answers at absolute best.
That said, it's not like nothing has changed. Alabama's offense has pulled a vanishing act. Clemson began the year with a three-touchdown loss to Duke. Notre Dame's offense might actually be explosive. Texas and Oklahoma are looking like how Texas and Oklahoma are always supposed to look (but rarely do at the same time). With the season about one-fourth done, then, it might be useful to reassess.
Let's revisit preseason expectations and see what has and hasn't shifted in some regard over the past three Saturdays.
Jump to a section:
Steady top 10
Time to panic?
Guess who's back
A stacked Pac-12
Biggest improvements
Possible top four
Almost nothing's changed for half the top 10

Georgia Bulldogs
Preseason SP+ ranking: No. 1, 10.6 average projected wins
After Week 3: No. 2, 10.6 average wins

Ohio State Buckeyes
Preseason SP+ ranking: No. 2, 10.0 average projected wins
After Week 3: No. 1, 10.1 average wins (+0.1)

Michigan Wolverines
Preseason SP+ ranking: No. 3, 10.4 average projected wins
After Week 3: No. 3, 10.4 average wins
Aside from Georgia and Ohio State trading places atop the SP+ rankings, the top three teams are almost exactly where they were three weeks ago. Michigan has looked far less convincing against its early season cupcakes than in previous years. The Wolverines outscored East Carolina, UNLV and Bowling Green by a combined 96-16, which sounds fine but is 46.2 points under what SP+ projected the combined margin to be. It also pales in comparison to the domination of the past couple of College Football Playoff years: They outscored three similar opponents 166-17 in 2022, and in 2021 they obliterated two of the MAC's better teams (WMU and NIU) and Washington by a combined 141-34.
The bar's high for proving you're at a playoff standard, and Michigan hasn't cleared it yet. Of course, none of this will matter if it pummels Rutgers on Saturday and starts looking the part.
It's a similar story for Georgia: The Dawgs have been under the projections by a combined 29.4 points, and in their first game against a power conference opponent (South Carolina), they laid a first-half egg and had to rally from an 11-point halftime deficit to win by 10. This is an extremely banged-up team, and maybe some semblance of good health will paper over all cracks. But there are a few cracks at the moment.
As with Michigan, this doesn't have to matter at all -- it hasn't changed the Bulldogs' win projections -- but it's still not to its usual standard. And Ohio State was pulling a similar act until Saturday's 63-10 demolition of Western Kentucky. One game doesn't necessarily mean a full trajectory shift, but it was awfully impressive.

Penn State Nittany Lions
Preseason SP+ ranking: No. 6, 9.6 average projected wins
After Week 3: No. 9, 9.9 average wins (+0.3)

USC Trojans
Preseason SP+ ranking: No. 10, 9.4 average projected wins
After Week 3: No. 5, 9.6 average wins (+0.2)
For two other aspiring contenders, things are going as planned. USC beat a trio of overwhelmed opponents (San Jose State, Nevada, Stanford) by a combined 178-52 and has climbed to fifth in both SP+ and the AP poll. After a shaky performance against SJSU, the Trojans' defense, so justifiably maligned a year ago, played well. They're currently 35th in points allowed per drive and 23rd in success rate allowed, and while those aren't elite numbers considering the competition levels, it's more than good enough to pair with what SP+ says is still the best offense in the country. The run defense is still questionable, but the pass rush and secondary have done their part.
Penn State looked outstanding in its first two games, then kept Illinois at arm's length in its first road trip of the season. The Nittany Lions fell in SP+ after the Illinois win because the "average fewer yards per play than Toledo did against Illinois, but force five turnovers" recipe isn't the most sustainable in the world. But for now we'll write it off because it was sophomore quarterback Drew Allar's first road start. He wasn't great (16-for-33, 208 yards), but he survived. He'll get a much bigger test at home against the top-ranked Iowa defense this Saturday.
On the verge of panic

Alabama Crimson Tide
Preseason SP+ ranking: No. 4, 9.6 average projected wins
After Week 3: No. 6, 8.5 average wins (-1.1)
It might be the most jarring 1% chance in the history of 1% chances. Alabama has won at least 10 games for 15 consecutive seasons, with 15 straight AP top-10 finishes, 14 straight SP+ top-three finishes, six national titles and three other title-game appearances. And one-quarter of the way through 2023, there's a 1% chance that Nick Saban's Crimson Tide go 5-7.
Granted, there's also a 3% chance that they rebound to win the national title. This is still Alabama we're talking about, and after the greatest 15-year run in the sport's history, the Tide will continue to get the benefit of the doubt from the numbers for a while longer. (Hence why they're still sixth overall.) But 5-7 is now the floor, and we got an extended look at that floor Saturday in Tampa. In a 17-3 win over USF, Bama quarterbacks Tyler Buchner and Ty Simpson combined to go 10-for-23 for 107 yards with five sacks and a Total QBR of 18.5 -- the worst for Bama QBs since a 20-6 win over South Carolina in 2009.
The Crimson Tide still won, thanks to a minimally challenged defense and 203 combined rushing yards from Roydell Williams and Jase McClellan. But the offensive line is an absolute disaster, and after Saban benched QB Jalen Milroe following a shaky performance in the loss to Texas, he found out that his backups were even shakier. Milroe will evidently start moving forward, but it's clear that what we thought was a questionable Alabama QB room is something far worse than questionable.
Granted, Bama won the national title in 2009, the last time it had a QB performance this poor. That year, the Tide grounded out wins with scores like 22-3, 12-10 and 24-15, then hit the accelerator in the SEC championship. Maybe defense and solid rushing will get them through upcoming fights with Ole Miss, Texas A&M and Tennessee, but they've underachieved against projections by 22.8 points in back-to-back games. Keep that up, and more losses are on the table (even if 5-7 probably isn't).

Clemson Tigers
Preseason SP+ ranking: No. 5, 9.7 average projected wins
After Week 3: No. 12, 8.3 average wins (-1.4)
Alabama isn't the only recent national champion to underachieve. Clemson's reputation is still reeling from the Tigers' 28-7 Labor Day loss at Duke in which the Tigers imploded in the red zone, suffered three turnovers and fell apart down the stretch.
They've been fine since, beating Charleston Southern and FAU by a combined 114-31, but they suffered another three turnovers in the process -- only 19 FBS teams have committed more through three weeks -- and to date, the defense has allowed 28.3 more points than SP+ projected.
FSU's near-stumble at Boston College actually dropped the Seminoles back down below the Tigers in SP+. Maybe there's a chance Clemson scores a big Saturday win and positions itself to win yet another ACC title, but despite offensive coordinator Garrett Riley coming to town to theoretically modernize the attack, quarterback Cade Klubnik is averaging a ghastly 9.8 yards per completion and ranks 95th in Total QBR. (DJ Uiagalelei during Clemson's 2021 offensive collapse: 97th.)
Dabo Swinney seemed to be making a big statement in hiring Riley, but instead Clemson is becoming less explosive and seemingly less capable of producing high-level QB play. The Tigers will absolutely need some chunk plays and big moments from Klubnik to beat the Noles on Saturday.

LSU Tigers
Preseason SP+ ranking: No. 7, 8.9 average projected wins
After Week 3: No. 15, 8.0 average wins (-0.9)

Tennessee Volunteers
Preseason SP+ ranking: No. 8, 8.9 average projected wins
After Week 3: No. 18, 7.9 average wins (-1.0)
It's possible that LSU is already on the way toward a rebound. Following a second-half collapse against Florida State and an unsteady start against Grambling, the Tigers pulled away to score 72 points against Grambling, then destroyed Mississippi State 41-14 in Starkville.
Jayden Daniels is back up to 18th in Total QBR, and LSU ranks fourth in points per drive. Granted, the defense still has to prove itself -- the Tigers are 82nd in points allowed per drive and 104th in QBR allowed -- and road trips to unbeaten Ole Miss and Missouri loom after this weekend's visit from Arkansas. Still, it looks like the recovery is already underway.
We'll see if that's the case for Tennessee. The Vols battled offensive droughts in otherwise comfortable wins over Virginia and Austin Peay, then lost their unbeaten record because of another drought. After going up 7-0 early against Florida, the Vols scored just three points in seven possessions and gave up a 29-3 run in an eventual 13-point loss. Joe Milton III threw a couple of pretty deep balls, and the Vols had a couple of questionable officiating decisions go against them, so maybe it wasn't as bad as it seemed. But SP+ projected them to score 90.7 points over the last two games, and they managed just 46. Their next two games are against UTSA (77th in defensive SP+) and South Carolina (62nd), and Milton better find his rhythm. After that, they take on five top-20 defenses in a six-game span.
A great start and a new lease on life

Texas Longhorns
Preseason SP+ ranking: No. 9, 9.1 average projected wins
After Week 3: No. 10, 10.0 average wins (+0.9)

Florida State Seminoles
Preseason SP+ ranking: No. 15, 8.9 average projected wins
After Week 3: No. 13, 9.6 average wins (+0.7)
Two of the biggest winners of the first quarter of the season were Florida State and Texas. The Seminoles played a nearly perfect second half to turn a halftime deficit into a resounding 21-point win over LSU in Week 1, while the Longhorns scored their biggest true road win since 1969 with their victory over Alabama in Week 2.
Both have otherwise left mixed impressions. FSU destroyed Southern Miss but fell victim to injuries, wind and a feisty opponent in narrowly avoiding a shocking upset loss to Boston College. Texas, meanwhile, underachieved versus projections in still-comfortable wins over Rice and Wyoming. Still, both have improved their projected win totals by a healthy amount. And FSU's will go through the roof this coming Saturday if the Seminoles can score their first win at Clemson in 10 years.

Oklahoma Sooners
Preseason SP+ ranking: No. 13, 9.2 average projected wins
After Week 3: No. 7, 10.2 average wins (+1.0)
Three other teams have improved their projections by similar amounts without the benefit of a marquee win. Oklahoma embarrassed Arkansas State and Tulsa and survived a stiff test from SMU in moving to 3-0 and jumping to seventh in SP+. We don't yet know how the Sooners will operate in close games (they were 0-5 in one-score finishes last season), but we at least got a hint in the way they kept the Mustangs at arm's reach and won by 17. They play at Cincinnati this coming Saturday, which could be awkward -- the Bearcats looked excellent in their first two games until a series of red zone mishaps led to an upset loss to Miami (Ohio) -- but OU's first huge test will come in three weeks against Texas.

Notre Dame Fighting Irish
Preseason SP+ ranking: No. 11, 8.8 average projected wins
After Week 3: No. 4, 9.9 average wins (+1.1)

Ole Miss Rebels
Preseason SP+ ranking: No. 17, 7.6 average projected wins
After Week 3: No. 14, 8.7 average wins (+1.1)
Notre Dame and Ole Miss don't have to wait that long. Whatever measure you choose -- points per drive (seventh on offense, 19th on defense), yards per play (fifth and third, respectively), success rate (eighth and 13th) -- the Fighting Irish have been excellent in 2023. They struggled for a bit against the only top-100 opponent they've played (NC State), but they still pulled away for a three-touchdown road win. They're both making and preventing big plays at a level we haven't seen from Notre Dame in a while. On Saturday, they will welcome to town what both SP+ and the FPI think is the best team in the country (Ohio State). If this is a genuinely different Notre Dame team, we'll know soon enough.
Ole Miss, meanwhile, gets another shot at Alabama this weekend. The Rebels have overachieved against SP+ projections by an average of 15.4 points per game thus far and pulled away to beat a talented Tulane team on the road in Week 2. They've won only twice ever in Tuscaloosa, but they did so with an explosive and free-wheeling team in 2015, and both of those adjectives apply to this Rebels squad. They're averaging 8.2 yards per play (eighth in FBS), and they've already posted 30 gains of 20-plus yards (second).
Goodness, the Pac-12 is loaded

Washington Huskies
Preseason SP+ ranking: No. 18, 8.7 average projected wins
After Week 3: No. 8, 9.8 average wins (+1.1)

Oregon Ducks
Preseason SP+ ranking: No. 12, 9.2 average projected wins
After Week 3: No. 11, 9.7 average wins (+0.5)

Utah Utes
Preseason SP+ ranking: No. 14, 8.5 average projected wins
After Week 3: No. 17, 8.7 average wins (+0.2)
It has been a funky season in the power conference universe. Four of the five conferences have seen their average projected win totals sink at least slightly, led by the SEC (minus-0.17 per team) and Big 12 (minus-0.15). But despite a couple of lead weights -- Stanford and Arizona State have gone from a combined 7.1 projected wins to 4.5 -- the Pac-12's average rose by 0.39 wins per team.
USC has lived up to its potential thus far, Colorado has exceeded projections, unbeaten Washington State and UCLA have gone from a combined 14.1 projected wins to 16.0, and despite each taking on power conference road tests, the trio of Utah, Oregon and Washington has gone a combined 9-0. Utah's offense remains wobbly as it waits for quarterback Cam Rising's return, but the Utes still beat both Florida and Baylor. Meanwhile, Oregon and especially Washington have shined.
Eight of the conference's 12 teams remain unbeaten through Week 3 -- the other nine FBS conferences average only 3.3 unbeatens each -- and the coming weeks are going to be one steady battle royale, starting with Colorado-Oregon, UCLA-Utah and Oregon State-Washington State this coming weekend.
Most improved win totals

Colorado Buffaloes
Preseason SP+ ranking: No. 88, 2.8 average projected wins
After Week 3: No. 76, 5.3 average wins (+2.5)

Texas State Bobcats
Preseason SP+ ranking: No. 122, 3.2 average projected wins
After Week 3: No. 95, 6.1 average wins (+2.9)
It's difficult not to notice that the two teams that have improved their projected win totals the most this season are both led by first-year coaches -- Colorado's Deion Sanders, Texas State's G.J. Kinne -- who loaded up on transfers more than anyone else. Heavy transfer portal usage doesn't guarantee immediate success (just ask Arizona State's Kenny Dillingham), but it certainly seems to broaden the range of potential outcomes.
We're still figuring out the Buffaloes' and Bobcats' respective ceilings. Colorado overachieved SP+ projections by a combined 59.3 points in wins over TCU and Nebraska, but underachieved in a narrow rivalry game win over Colorado State. We'll see if that was a speed bump or a sign that they're approaching their ceiling. I'm guessing the former, but either way, the schedule is also about to get much more difficult with games against Oregon and USC in the next two weeks.
Texas State, on the other hand, overachieved by a combined 90.9 points in wins over Baylor and Jackson State and a narrow loss to UTSA. The Bobcats are now projected favorites in five of their final nine games, and SP+ still potentially hasn't caught up to how good they are.

Duke Blue Devils
Preseason SP+ ranking: No. 49, 6.1 average projected wins
After Week 3: No. 32, 8.2 average wins (+2.1)
It took some red zone funkiness for Duke to beat Clemson by 21 points in Week 1, but make no mistake: It wasn't funky that the Blue Devils won. And since then, they have handled their business as projected against Lafayette and Northwestern. They are up to 32nd in SP+ -- they haven't finished a season that high since 2014 -- and they are positioning themselves to match or exceed their 8-4 record last regular season. Mike Elko has earned loads of plaudits for his work in immediately turning this program back around (and with minimal transfer portal help), but we probably still aren't talking about it enough.

Rutgers Scarlet Knights
Preseason SP+ ranking: No. 78, 4.0 average projected wins
After Week 3: No. 64, 6.0 average wins (+2.0)
This is the third straight year that Rutgers has started 3-0; they went a combined 3-16 thereafter in 2021 and '22. Maybe a similar fate awaits, but it's noteworthy that Greg Schiano's Scarlet Knights have overachieved SP+ projections by double digits in every game. They're up to 64th, and their odds of finishing bowl-eligible for the first time since 2014 are up to 65.9%, but they might be nowhere near their ceiling yet.

FIU Panthers
Preseason SP+ ranking: No. 130, 4.3 average projected wins
After Week 3: No. 106, 6.3 average wins (+2.0)
The Golden Panthers have finished above .500 four times in 19 FBS seasons; this is a pretty hard job. But Mike MacIntyre flipped SJSU from 1-12 to 10-2 and flipped Colorado from 2-10 to 10-4 within two years each. He isn't afraid of a challenge. He inherited an FIU program that had lost 18 of 19 games, and in Year 2, his team is 3-1 and has a 71.7% chance of reaching bowl eligibility. He is a turnaround artist.
The quarter ahead
Last year I stumbled across something very useful. In an effort to create a BCS-like formula for hypothetical use in the national title race, I instead ended up creating a formula that rather accurately mimics how the CFP committee ranks teams. Call it the gooey butter cake of formulas -- it wasn't what was intended, but it was delicious all the same.
It's too early to be paying serious attention to the national title race, but I figure knowing (somewhat) how the playoff committee thinks is still pretty useful when it comes to framing the coming weeks. When are the biggest matchups? Who's got the most at stake between now and the season's midway point?
Using this accidental formula -- which basically amounts to two-thirds poll rankings, one-third résumé rankings from SP+ and the FPI -- here's what the faux CFP top 25 looks like three weeks into the season:
1. Georgia (3-0)
2. Michigan (3-0)
3. Florida State (3-0)
4. Texas (3-0)
5. Ohio State (3-0)
6. USC (3-0)
7. Penn State (3-0)
8. Washington (3-0)
9. Notre Dame (4-0)
10. Oregon (3-0)
11. Utah (3-0)
12. LSU (2-1)
13. Alabama (2-1)
14. Oklahoma (3-0)
15. Oregon State (3-0)
16. Ole Miss (3-0)
17. North Carolina (3-0)
18. Duke (3-0)
19. Miami (3-0)
20. Colorado (3-0)
21. Tennessee (2-1)
22. UCLA (3-0)
23. Washington State (3-0)
24. Iowa (3-0)
25. Clemson (2-1)
Building from this list, we've got 17 ranked-versus-ranked matchups between now and Week 7. A whopping seven of those games happen this coming Saturday. This is legitimately one of the biggest September weekends in recent college football history.
Week 4: No. 5 Ohio State at No. 9 Notre Dame, No. 3 Florida State at No. 25 Clemson, No. 16 Ole Miss at No. 13 Alabama, No. 20 Colorado at No. 10 Oregon, No. 24 Iowa at No. 7 Penn State, No. 22 UCLA at No. 11 Utah, No. 15 Oregon State at No. 23 Washington State
Week 5: No. 6 USC at No. 20 Colorado, No. 11 Utah at No. 15 Oregon State, No. 9 Notre Dame at No. 18 Duke, No. 12 LSU at No. 16 Ole Miss
Week 6: No. 4 Texas vs. No. 14 Oklahoma, No. 23 Washington State at No. 22 UCLA
Week 7: No. 6 USC at No. 9 Notre Dame, No. 10 Oregon at No. 8 Washington, No. 18 Miami at No. 17 North Carolina, No. 22 UCLA at No. 15 Oregon State
We still aren't going to learn much about the top two teams, Georgia and Michigan. But teams like Notre Dame and a number of aspiring Pac-12 teams are facing an enormous month ahead.