September is always a dizzying college football experience. After the world's longest offseason, we are desperate to overreact to everything we see. We careen from wild finish to wild finish and from breathless Saturday to breathless Saturday. And then, as the calendar flips to October ... we realize not a lot has changed in terms of the big picture.
The final Saturday of September gave us another round of thrills and once again made it really hard for me to contain the My 10 Favorite Games list at the bottom of this column. Georgia pulled off its greatest comeback in school history in the biggest game of the week, only to lose anyway. Miami "lost" via a last-second Hail Mary but won anyway. Oklahoma rose from the dead. Long field goals and awful field goals decided countless games.
And yet, as we head into October, we see that the top nine teams in the preseason AP poll are a combined 36-4, with two of four losses coming to other top-nine teams. Ohio State is an even bigger Big Ten favorite than it was a month ago. Bama and Georgia are both still likely to make the College Football Playoff.
As fun as the first five weeks of the season have been -- and good lord, they've been awesome -- has anything really changed?
Yes, actually! You just have to know where to look. Here are some of 2024's most interesting plot twists.
Jump to a section:
Georgia, Michigan slip | Miami rises
Big 12 intrigue | Navy, Army are legit
Auburn's anguish | Flags are flying
Week's biggest surprises
Heisman of week | 10 favorite games

Recent national champs have stumbled
Heading toward Saturday, I saw plenty of grumbling on social media about how the buzz for Week 5's Georgia-Alabama game was lower than such a game should be, at least in part because of the expanded College Football Playoff and the fact that both teams were likely to make the CFP no matter who won.
Now, some of this stems from the way we're able to fool ourselves on social media -- the "WHY ISN'T ANYONE TALKING ABOUT [thing lots of people are talking about]???" effect. There was all sorts of Georgia-Bama content atop the ESPN college football page throughout the week, and people clicked on it. ESPN's "College GameDay" was in Tuscaloosa, and when the TV ratings come in, I'm pretty sure they're going to show that both "GameDay" and the game itself had a metric crap-ton (technical term) of viewers who were rewarded with 2024's game of the year to date, a 41-34 Bama win.
There's no question the stakes were lower in Tuscaloosa than they have been in the past. Granted, it was less than three years ago when Georgia lost to Alabama (in the 2021 SEC championship game) and went on to win the national title anyway, so we shouldn't lie to ourselves and suggest this game would have been do-or-die in some previous universe. But after losing, Georgia's odds of making the CFP are still at 72.3%, per the Allstate Playoff Predictor.
The game was still absolutely incredible, however. If someone with no knowledge of how college football's title is determined watched the last three minutes of the fourth quarter -- the long touchdowns, the player celebrations, the fan expressions -- that person would have thought this game was for the national title. No one in or around Bryant-Denny Stadium wanted that game less because of playoff expansion.
Second, even if the stakes of Bama-Georgia were a little lower than they could have been with a smaller playoff (or no playoff at all), games like Kansas State-Oklahoma State, Penn State-Illinois, Boise State-Washington State and UNLV-Fresno State all had bigger stakes because of the playoff's expansion. For every "means less" game the rest of the way, there will be quite a few "means more" games.
Finally, there were still stakes in Tuscaloosa! Alabama is now quite a bit more likely to win the SEC title and score a first-round bye, and Georgia has a hole to dig itself out of moving forward. So does Michigan.
The 10 teams with the largest regression in SP+ ratings since the preseason:
1. Florida State -13.3 adjusted points per game (ranking fell from 12th to 48th)
2. UCLA -12.6 (from 37th to 84th)
3. UTSA -11.2 (from 54th to 100th)
4. Michigan -9.3 (from seventh to 15th)
5. Wisconsin -9.3 (from 26th to 66th)
6. Georgia -8.8 (from first to seventh)
7. Arkansas State -8.3 (from 93rd to 123rd)
8. Kennesaw State -8.2 (from 131st to 134th)
9. MTSU -7.7 (from 114th to 129th)
10. NC State -7.5 (from 29th to 61st)
Over seven previous seasons, Georgia only once finished lower than fifth in SP+. Not only are the Dawgs currently seventh, but they must still play three teams ranked higher than them, two on the road. If the Dawgs don't start to improve back toward the top-ranked team they were supposed to be, they're going to be begging the playoff committee to let them in at 9-3. (If they even get to 9-3.)
Michigan's struggles were perhaps more predictable: The Wolverines lost coach Jim Harbaugh and almost the entire starting offense after last season's national title run, and they headed into the season without anyone who could pass reliably. They haven't yet played on the road, but they've already lost one blowout to Texas and nearly lost games to USC (22nd in SP+) and, on Saturday, Minnesota (47th). They have yet to live up to SP+ projections in five games and could fall even further in the weeks to come. Even if we expected regression, we didn't quite expect this.
Miami has replaced Florida State
In the intro, I noted how well the preseason AP top nine is doing. I had to do that primarily because the preseason No. 10 team, Florida State, has collapsed in just about every conceivable way. The Seminoles were preseason ACC favorites, beginning the year with a 36% chance of making the CFP, per the Allstate Playoff Predictor.
Following Saturday's latest debacle, a 42-16 humiliation at the hands of ACC newcomer SMU, the Seminoles are 1-4. In their lone win, a 14-9 defeat of Cal, was statistically unlikely -- they were outgained by 126 yards and allowed far more scoring opportunities than they created. Almost halfway through the season, they're 125th in total offense and 92nd in total defense. The only thing they've been good at is keeping the pace low enough that their defense doesn't get torched even more. When it comes to preseason expectations, no one can even come close to FSU's level of underachievement.
The 10 teams with the largest drop in CFP odds since the preseason
1. Florida State -36.2% (from 36.2% to 0.0%)
2. Oklahoma -26.7% (from 36.5% to 9.8%)
3. Kansas -24.5% (from 24.6% to 0.1%)
4. Michigan -19.1% (from 28.1% to 9.0%)
5. Oregon -15.3% (from 74.8% to 59.5%)
6. UTSA -14.8% (from 14.8% to 0.0%)
7. NC State -13.7% (from 13.8% to 0.1%)
8. Auburn -13.3% (from 13.3% to 0.0%)
9. Notre Dame -12.4% (from 58.8% to 46.4%)
10. LSU -14.1% (from 25.4% to 14.1%)
Outside the state of Florida, the season has more or less played out as it was supposed to in the ACC. Louisville's conference title odds are up a bit, and NC State's own ridiculously disappointing start has eliminated the Wolfpack from the conversation. In fact, if we just pretend we were hyping a different Florida team in the offseason, things look downright orderly. Because while Florida State has underachieved significantly, Miami has improved its own odds by similar margins.
The 10 teams with the largest rise in CFP odds since the preseason
1. Miami +43.9% (from 17.4% to 61.3%)
2. Tennessee +42.6% (from 36.4% to 79.0%)
3. Indiana +39.0% (from 0.6% to 39.6%)
4. Alabama +33.9% (from 57.2% to 91.1%)
5. UNLV +29.5% (from 1.0% to 30.5%)
6. Texas +23.2% (from 67.9% to 91.1%)
7. BYU +23.0% (from 0.7% to 23.7%)
8. Boise State +22.5% (from 15.4% to 37.9%)
9. Ole Miss +20.7% (from 20.4% to 41.1%)
10. Iowa State +20.4% (from 4.7% to 25.1%)
After underachieving significantly over the past couple of years (OK, decades), the Hurricanes have bolted out of the gates in 2024. They even survived their first gut-check moment of the season, overcoming a poor start and late chaos to beat Virginia Tech. Road trips to Cal and Louisville in the next two weeks could lead to a crash, but through September they're the real deal.
The Big 12 remains a glorious mystery
Through five weeks, five teams, all in the Group of 5, have seen their conference title odds, per SP+, improve by at least 20 percentage points (Army and Navy in the AAC, James Madison in the Sun Belt, UNLV in the Mountain West and Toledo in the MAC) and five have seen their odds decrease by that much (Georgia in the SEC, Florida State in the ACC, Memphis and UTSA in the AAC and Miami-Ohio in the MAC). These are pretty big shifts considering we haven't seen a ton of conference games yet.
The Big 12 doesn't have a single team that has seen its odds surge or plummet like that. Instead, we've seen almost weekly surges and crashes. In the past two weeks alone, Kansas State's title odds have gone from 28.0% to 11.6% following a strange blowout loss to BYU, then jumped back up to 23.3% following a blowout of Oklahoma State. Utah looked like a major front-runner after four weeks (36.0%), but after an offensively dire home loss to Arizona, the Utes are now the No. 3 favorite (15.1%).
Current Big 12 title odds, per SP+:
Kansas State 23.3% (up 1.8% from preseason)
Iowa State 20.2% (up 12.5%)
Utah 15.1% (down 5.3%)
BYU 8.7% (up 8.1%)
Arizona 6.9% (down 4.0%)
TCU 6.0% (up 1.1%)
Colorado 5.8% (up 4.3%)
UCF 4.2% (up 1.3%)
Texas Tech 3.5% (down 0.2%)
West Virginia 2.6% (down 2.5%) Cincinnati 1.4% (up 0.6%)
Arizona State 1.0% (up 0.4%)
Oklahoma State 0.8% (down 10.7%)
Kansas 0.5% (down 5.8%)
Baylor 0.1% (down 1.3%)
Houston 0.0% (down 0.3%)
Iowa State and BYU have pulled little-engine-that-could routines, slowly climbing the hill with lovely defense and, more importantly, zero losses. But so many other teams have charged and crashed only to end up close to where they started five weeks ago. It's still nearly nine weeks until Championship Week. We've got time for about 400 more plot twists here.
Army and Navy are looking ridiculously good
The top of the current Total QBR rankings makes at least a certain amount of sense. Jalen Milroe and Cam Ward, aka the top two current Heisman betting favorites per ESPN BET, are fifth and second, respectively, and the QBs for plenty of other top teams -- Penn State's Drew Allar (sixth), Ole Miss' Jaxson Dart (seventh), Texas' Arch Manning (eighth), Ohio State's Will Howard (12th), Georgia's Carson Beck (16th), Oregon's Dillon Gabriel (18th) -- are pretty high up there as well, even if some, like Beck and Gabriel, have had more disappointing moments than expected.
No. 1 on the list, however, is Navy's Blake Horvath. Not even Midshipmen fans would have predicted that one. Horvath is crafting new levels of option brilliance: In four games, he has rushed for 470 yards (not including three sacks), and while he has taken three sacks and thrown an interception among his 49 dropbacks, he also has completed 30 of 44 passes at 21.2 yards per completion. So that's 10.7 yards per (non-sack) rush and 12.9 yards per dropback.
You know what's even scarier? He's getting better: In the past two games, Horvath has averaged 13.9 yards per rush and 16.2 yards per dropback. That should only be possible in video games.
Army's Bryson Daily, meanwhile, might pale in comparison to Horvath's recent level, but he's 36th in Total QBR, ahead of Missouri's Brady Cook and Colorado's Shedeur Sanders, among others. He completed two passes for 54 yards while rushing for 152 yards and three scores in a Thursday night blowout of Temple.
Horvath and Daily have led Army and Navy to matching 4-0 starts, and they currently boast matching 31.3% chances of winning the AAC, per SP+. In the transfer portal era, coaches can change the DNA of their team in record time and, like Curt Cignetti and his merry band of transfers at Indiana, give themselves a chance to dramatically overachieve against preseason projections. But the biggest and fourth-biggest overachievers of the season to date took in a combined zero transfers.
The 10 teams with the largest improvement in SP+ rating since the preseason:
1. Navy +16.5 adjusted points per game (ranking rose from 113th to 68th)
2. Indiana +15.8 (from 81st to 29th)
3. UConn +14.8 (from 125th to 81st)
4. Army +13.8 (from 103rd to 70th)
5. Pitt +11.1 (from 66th to 31st)
6. UNLV +10.7 (from 82nd to 54th)
7. Vanderbilt +10.3 (from 89th to 69th)
8. San José State +10.0 (from 101st to 75th)
9. BYU +9.4 (from 69th to 42nd)
10. James Madison +8.8 (from 56th to 33rd)
The Midshipmen and Black Knights destroyed conference foes by a combined 83-32 in Week 5. They've refreshed and renewed their option attacks, and both are playing some downright sexy ball at the moment. There's a long way to go, but we should probably get used to the idea of potentially having back-to-back Army-Navy games -- first in the AAC championship, then in the normal Army-Navy slot -- late this fall. There might be some CFP connotations to those games as well.
Dillon vs. Carson vs. Quinn? No, no. Jalen vs. Cam vs. Travis
All things considered, the Heisman race hasn't really changed much to date. Among the top 10 current betting favorites, per ESPN BET, seven were in the preseason top 10 back in July, including each of the preseason top four (Oregon's Dillon Gabriel, Georgia's Carson Beck, Texas' Quinn Ewers and Ole Miss' Jaxson Dart).
Those four guys are secondary favorites at the moment, however, all with odds between +1600 and +3000. The current top four consists of two guys who have jumped from secondary to primary (Jalen Milroe and Cam Ward) and two who weren't among that preseason top 10 (Travis Hunter and Ashton Jeanty).
Current Heisman odds, per ESPN BET
1. Jalen Milroe, Alabama (+225)
2. Cam Ward, Miami (+600)
3. Travis Hunter, Colorado (+700)
4. Ashton Jeanty, Boise State (+1100)
5T. Dillon Gabriel (+1600)
5T. Nico Iamaleava, Tennessee (+1600)
7. Jaxson Dart (+2200)
8. Quinn Ewers (+2500)
9T. Carson Beck (+3000)
9T. Miller Moss, USC (+3000)
Milroe's stock obviously rose considerably following his brilliant performance against Georgia, but Ward has been the most consistently Heisman-worthy quarterback of the season, Jeanty is on pace for 3,000 yards from scrimmage and more than 40 touchdowns (projected over 14 games), and Hunter just becomes more Hunter by the day: He's currently eighth nationally in receiving yards and tied for 14th in interceptions.
Past experience has taught us that September's Heisman favorites aren't December's Heisman winners. But the top four have earned their places in the race, and if this is the race we continue to get, it will be as entertaining as any in recent memory.
Auburn is making history (and not in the good way)
I have a measure called postgame win expectancy, or PGWE for short, which looks at the stats from a given game (the things that end up feeding into SP+), tosses them into the air and declares, "With these stats, Team A could have expected to win this game X% of the time."
To date, there have been 13 games in which a team with a 75% or greater postgame win expectancy actually lost. Auburn lost two of them in eight days. First came last week's masterpiece, in which the Tigers produced a 93.3% PGWE against Arkansas but fell 24-14 thanks to five turnovers, three inside the Arkansas 30. It was the second time this season that they've committed five turnovers (along with the loss to Cal), and it was the second game they might have actually won had they committed only three!
By comparison, Saturday's 27-21 loss to Oklahoma was downright orderly; it featured only a 75.3% PGWE! The Tigers outgained the Sooners by 191 yards and more than doubled their first downs (26-11), but they got stuffed on fourth-and-goal from the 1 in the first quarter, Towns McGough missed field goals of 27 and 51 yards, and while the Tigers managed to commit only one turnover, it was a 63-yard pick-six as they were driving to ice the game.
If you add up a team's PGWE for each game, you get a general look at what your average win total should be considering the stats you produce. Each year produces over- and underachievers in this regard, and this year is no different: PGWE suggests, for instance, that Minnesota should have about 3.4 wins (a 3-2 or 4-1 record), but the Gophers are 2-3. Washington should have 4.5 wins but is 3-2. Kansas should have 2.7 wins but is 1-4.
As we turn to October, Auburn is your leading underachiever. The Tigers' first five games produced stats that suggest an average of 3.9 wins; they were far more likely to go 4-1 than even 3-2. But they're 2-3. Last year's biggest PGWE underachievers, TCU and Nebraska, each finished the season 2.0 wins short of their expected record. Auburn has nearly matched that in less than half a season. History in the making!
It's not your imagination: There are more penalties
In 2023, officials called 5.82 penalties per team per game for an average of 51.1 yards. As you might expect, these averages were higher earlier in the season, as teams were perhaps working out the kinks and building a rhythm: Before October, teams averaged 5.99 penalties and 52.7 penalty yards per game.
This year, teams are averaging 6.38 penalties and 56.4 penalty yards per game.
Week 5 was actually the most lawful weekend since Labor Day, but 10 still teams hit 90 or more penalty yards and 32 teams committed at least eight penalties. Purdue and Nebraska combined for 24 penalties and 260 yards. Auburn and Oklahoma combined for 17 and 139, but 12 of those penalties came in a second half that grew absolutely dire before some late-game fireworks.
I'm not sure who asked for more whistles, but we're getting them.
The five most surprising results of Week 5
Here are the five results that were furthest away from their respective SP+ projections. Call them either surprises or bad projections, I guess.
1. UNLV 59, Fresno State 14 (projection: UNLV by 2.6). To replace USC-bound quarterback Jayden Maiava last offseason, UNLV head coach Barry Odom decided two FCS transfers were better than one. He landed Holy Cross' Matthew Sluka, whom you've undoubtedly heard of by now. But he only narrowly won the job over former Campbell QB Hajj-Malik Williams. Williams eased into the starting lineup Saturday and played brilliantly efficient ball while the Rebels' defense mostly dominated and their -special teams scored twice. This was an incredible performance by what has to be considered one of two front-runners for the Group of 5's automatic CFP bid. (We'll get to the other one in a second.)
2. UConn 47, Buffalo 3 (projection: UConn by 2.7). Since a Week 1 throttling by Maryland dropped UConn to 132nd in SP+, the Huskies have ignited. They've overachieved against SP+ projections by an average of 31.0 points per game in four wins; it's like they suddenly became a top-30 team overnight.
3. Colorado 48, UCF 21 (projection: UCF by 12.3). This honestly might have been the most complete Colorado performance of the Deion Sanders era. The Buffaloes seemed to make every key play, got big plays from six or seven different playmakers, recorded 13 tackles for loss and might have ever-so-slightly entered themselves into the Big 12 title picture.
4. James Madison 63, Ball State 7 (projection: JMU by 26.4). This is the kind of game we completely ignore and SP+ pays a lot of attention to. JMU was supposed to blow out Ball State, but the Dukes utterly destroyed the Cardinals, treating them like a JV team. As a result, James Madison hopped to 33rd in SP+ and became the G5's CFP co-favorite with UNLV.
5. Tulane 45, South Florida 10 (projection: Tulane by 7.4). We'll see if Jon Sumrall's Green Wave can overcome nonconference losses to Kansas State and Oklahoma and insert themselves back into the G5 CFP picture. On Saturday, they reminded us that they have just about the highest upside of any G5 team.
Who won the Heisman this week?
I am once again awarding the Heisman Trophy every single week of the season and doling out weekly points, F1-style (in this case, 10 points for first place, 9 for second and so on). How will this Heisman race play out, and how different will the result be from the actual Heisman voting?
Here is this week's Heisman top 10:
1. Jalen Milroe, Alabama (27-for-33 passing for 374 yards, 2 TDs and 1 INT, plus 117 rushing yards and 2 scores against Georgia)
2. Ashton Jeanty, Boise State (26 carries for 259 yards and 4 TDs against Washington State)
3. Ryan Williams, Alabama (six catches for 177 yards and a touchdown against Georgia)
4. Travis Hunter, Colorado (nine catches for 89 yards and a touchdown, plus two tackles, an acrobatic interception and a pass breakup against UCF)
And I do mean acrobatic:
WHAT AN INTERCEPTION BY TRAVIS HUNTER
— FOX College Football (@CFBONFOX) September 28, 2024
He makes the play and hits the Heisman pose 🔥@CUBuffsFootball pic.twitter.com/cF3MLeOpuM
5. Cam Ward, Miami (24-for-38 passing for 343 yards, 4 TDs and 2 INTs, plus 73 rushing yards and a touchdown against Virginia Tech)
6. Caden Durham, LSU (seven carries for 128 yards and a touchdown, plus 89 receiving yards and a score against South Alabama)
7. Abdul Carter, Penn State (7 tackles, 4 TFLs, 2 sacks, a forced fumble and a pass breakup against Illinois)
8. Darian Mensah, Tulane (18-for-22 passing for 326 yards and 3 touchdowns against South Florida)
9. Donovan Ezeiruaku, Boston College (14 tackles, 4 TFLs, 3 sacks, a forced fumble and 2 QB hurries against Western Kentucky)
10. Blake Horvath, Navy (9-for-11 passing for 225 yards and 2 touchdowns, plus 84 rushing yards and a score against UAB)
Each of the top five players had a solid claim to No. 1. Milroe produced nearly 500 combined passing and rushing yards against the preseason No. 1 team, Ryan Williams had at least two of the best plays of the season from a receiver (and he's a 17-year old freshman), Jeanty did a number of incredible things (that I'm already starting to take for granted) in a vital win over Washington State, Hunter scored a TD and made one of those slithering, out-of-nowhere interceptions that have quickly become his calling card, and despite some early miscues, Ward led a fourth-quarter comeback and reminded everyone that he might be the best pure playmaker in college football.
What a week.
Honorable mention:
• Tahj Brooks, Texas Tech (32 carries for 172 yards and 2 touchdowns against Cincinnati)
• DJ Giddens, Kansas State (15 carries for 187 yards and a touchdown, plus 22 receiving yards against Oklahoma State)
• Eric O'Neill, James Madison (3 tackles, 2 TFLs, 2 sacks and a pick-six against Ball State)
• Shedeur Sanders, Colorado (28-for-35 passing for 290 yards, 3 TDs and 1 INT, plus 28 rushing yards against UCF)
• Nic Scourton, Texas A&M (4 tackles, all for loss, with 2 sacks, a forced fumble, a pass breakup and a QB hurry against Arkansas)
• Star Thomas, Duke (30 carries for 166 yards and a touchdown, plus 45 receiving yards and a score against North Carolina)
• Hajj-Malik Williams, UNLV (13-for-16 passing for 182 yards and 3 touchdowns, plus 119 rushing yards and a score against Fresno State)
Through five weeks, here are your points leaders:
1. Cam Ward, Miami (25 points)
2. Ashton Jeanty, Boise State (24 points)
3. Jalen Milroe, Alabama (18 points)
4. Travis Hunter, Colorado (17 points)
5. Jaxson Dart, Ole Miss (14 points)
6T. Alonza Barnett III, James Madison (10 points)
6T. Blake Horvath, Navy (10 points)
6T. Cade Klubnik, Clemson (10 points)
6T. Arch Manning, Texas (10 points)
10T. Antario Brown, Northern Illinois (9 points)
10T. Dillon Gabriel, Oregon (9 points)
10T. Miller Moss, USC (9 points)
I like that the top four are also the top four in the current betting odds. But congrats to Ward for becoming your esteemed September Heisman winner.
My 10 favorite games of the weekend
1. No. 4 Alabama 41, No. 2 Georgia 34.
Week 5 gave us three plays that impacted a game's win probability by at least 55%. Two of them -- Carson Beck's 67-yard pass to Dillon Bell that gave Georgia its first lead with 2:31 left, and Milroe's 75-yard bomb to Williams on the next play from scrimmage -- happened within 11 seconds of each other in Tuscaloosa. This game was interesting for the full 60 minutes, first because Bama was so thoroughly demolishing Georgia, then because the Dawgs kept pecking away and making things more and more tense. But the final three minutes took it into a different stratosphere. Considering the atmosphere, the teams' rankings and the number of unexpected plot twists, this was the game of the season to date.
2. No. 7 Miami 38, Virginia Tech 34 (Friday). For the second straight week, Friday gave us multiple classics, and in a different week this would have been an easy No. 1. Quarterback cousins Cam Ward (Miami) and Kyron Drones (Tech) combined for six passing touchdowns, and Tech, thanks in part to its first good first half of the season, held a 10-point lead with nine minutes left. Miami stormed back with help from a Ward-to-Riley Williams chest pass to lead in the final minute. Despite some wobbly clock management, the Hokies got a shot at the end zone and initially seemed to have completed a Hail Mary on the final play. But replay reversed it.
HAIL MARY NO GOOD 😳
— ESPN (@espn) September 28, 2024
No. 7 Miami outlasted Virginia Tech after the Hokies' final play was ruled incomplete. pic.twitter.com/Xrpl2UFod9
3. Sam Houston 40, Texas State 39. What an awesome turnaround for KC Keeler and Sam Houston. The Bearkats began their FBS life in 2023 with eight straight losses, but they've won seven of nine since and moved to 4-1 this season by scoring the game's last 19 points, including a 24-yard field goal with six seconds left, to turn a 39-21 deficit into an epic win.
4. FCS: Drake 30, San Diego 28. The stakes weren't quite as high in this Pioneer League battle as they were in Tuscaloosa, but Drake basically pulled a Bama in the in-game win probability department.
The Bulldogs watched an 18-point fourth-quarter lead vanish behind a 19-0 San Diego run. The Toreros took a shocking 28-27 lead with 1:55 left, but Shane Dunning's 48-yard field goal at the buzzer saved a chaotic Drake win.
5. Duke 21, North Carolina 20. I assumed that in response to Week 4's 70-50 loss to James Madison, North Carolina would either play its best game of the season or collapse further. The Tar Heels did both. They played an almost perfect first half and led 20-0 with six minutes remaining in the third quarter, but Duke scored three touchdowns in four drives to take a sudden lead with 5:42 remaining, and a Tre Freeman interception sealed a ferocious comeback win.
6. Rutgers 21, Washington 18 (Friday). Washington outgained Rutgers by 222 yards (2.7 per play), produced a 53% success rate to Rutgers' 39% and created more than twice as many scoring opportunities. That postgame win expectancy measure I mentioned above? Washington's was 96.7%. But the Huskies forgot to actually win. Rutgers scored touchdowns, Washington settled for (and missed) field goals and the Scarlet Knights snared the least likely win of Week 5.
7. Texas Tech 44, Cincinnati 41. Cincinnati led by 11 early. Texas Tech led by 14 late. There were touchdown passes of 42, 49 and 71 yards, plus a 51-yard pick-six. Cincinnati pulled within 3 points with 3:24 left and earned a 51-yard field goal attempt at the buzzer to force overtime. But Nathan Hawks missed it pretty badly, and Tech moved to 4-1.
8. No. 21 Oklahoma 27, Auburn 21. I noted above that there were three plays that altered a game's win probability by more than 55%, and two of them were in Bama-Georgia. The third? Kip Lewis' 63-yard pick-six that gave OU a stunning and sudden lead with 4:06 left at Auburn. The Tigers had avoided the types of disasters we saw in their first two losses and were on the brink of putting away a lovely (if sloppy and penalty-laden) home win. Instead, they fell to 2-3.
9. Louisiana 41, Wake Forest 38. College football's sheer depth is unbeatable. You couldn't be blamed for not tuning in to ESPN+ to watch Louisiana play a down-and-out (and soon to be 1-3) Wake Forest team in Winston-Salem, but if you did, you got to watch the Cajuns take (and lose) three separate leads, make a 31-yard field goal with 52 seconds left to take a fourth lead and watch Wake's Matthew Dennis miss a 42-yarder at the buzzer.
10. NAIA: No. 19 College of Idaho 45, No. 10 Montana Tech 37. Our friends out in Idaho had briefly fallen on hard times, losing two straight to fall from third in the NAIA rankings to nearly out of the top 20. The Yotes watched a 10-point homecoming lead turn into an 11-point fourth-quarter deficit on Saturday, too. But in a 44-point fourth quarter, they flipped a 31-19 deficit into a 37-31 lead with 1:53 left, and when Montana Tech drove 85 yards in less than a minute to score and tie the game at 37-37, they blocked the PAT to force overtime. Hunter Gilbert's 2-yard score and Dee'Shon Swafford's game-clinching INT gave the Yotes a vital (and somewhat delirious) win over the Orediggers.