"How did you go bankrupt?" Bill asked.
"Two ways," Mike said. "Gradually and then suddenly." -- Ernest Hemingway, "The Sun Also Rises"
In the present tense, it can seem impossible for a new team to join college football's ruling class. You fall short again and again, and it starts to feel like it's not going to happen. But then it just sort of does.
On the cusp of the 2015 season, we still had no idea what Dabo Swinney's Clemson Tigers might be capable of achieving. Swinney had gradually built a rock-solid program. The Tigers had won eight games per year from 2009 to 2011 with an average SP+ rating (presented as an adjusted points per game figure) of 11.5, then improved to 11.7 wins per year with an average rating of 20.4 from 2012 to '14.
They had steadily improved on defense since hiring coordinator Brent Venables, and the offense had enjoyed a strong run before freshman quarterback Deshaun Watson kept getting hurt in 2014. This was clearly a very good program, but it was in no way obvious that Clemson was about to become Clemson. They were ranked 12th in 2015's preseason AP poll.
Suddenly, the Tigers were elite. Hemingway's bankruptcy in reverse. They reached the CFP title game in 2015, then did so again in 2016, 2018 and 2019, winning two national titles. They're currently No. 1 in the AP poll and looking as dominant as ever.
Clemson's rise was a mere five years ago. LSU came out of nowhere last fall. Who might be next?
If we look at the average SP+ ratings from the past three seasons, the top six are probably who you'd imagine: Alabama, Ohio State and Clemson, plus 2019 champion LSU, 2017 runner-up Georgia and perennial Big 12 champion and CFP stalwart Oklahoma.
Plenty of other teams are above that Clemson line, though: Penn State averaged a 23.1 rating from 2017 to '19, Auburn 23.1, Wisconsin 20.9 and Michigan 20.8. Washington (20.1) and Notre Dame (19.7) are close too. (LSU, by the way, had a pre-breakthrough average of 23.5.)
We might not see the next breakthrough coming, but if it does, these programs are the most likely to make the jump. We'll use Penn State as a case study.
"We've gone from an average football team to a good football team to a great football team, but we're not an elite football team yet. ... Right now we're comfortable being great, and I'm gonna make sure that everybody in our program, including myself, is very uncomfortable. Because you only grow in life when you're uncomfortable." -- Penn State coach James Franklin, 2018
James Franklin took over a program in turmoil in 2014 and went just 7-6 in each of his first two seasons. With new offensive coordinator Joe Moorhead in place, the Nittany Lions surged to a surprising Big Ten title in 2016, finishing 11th in SP+. They haven't been able to get past Ohio State in the years since, going 0-3 against the Buckeyes and 31-5 against everyone else, but from a "first three years, next three years" perspective, the Nittany Lions are awfully close to Clemson's trajectory.
2014-16: 8.3 wins per year, 13.2 average SP+ rating (26.3 average ranking)
2017-19: 10.3 wins per year, 23.1 average SP+ rating (9.0 average ranking)
They begin their 2020 season Saturday eighth in the AP poll and fifth in SP+. Their upside for this fall was dampened by the losses of linebacker Micah Parsons to an opt-out and Journey Brown to unspecified health concerns, but in terms of overall program health and trajectory, you'd rather be Penn State than almost any other program in the FBS.
What separates the sport's current ruling class from the Penn States of the world? I asked PSU coaches that question this past summer.
Franklin's mind went immediately to offense. "The teams that are able to separate themselves consistently are the ones that can score points in a lot of different ways," he said. "In today's football, as good as you can be on defense, there's gonna be a couple of games a year where you're gonna have to outscore people. You're gonna have to score 40-something points in a big-time game to get a W."
If that's an exaggeration, it's barely one. Ohio State beat PSU 39-38 in 2017, and even with a defense that ranked 11th or better in defensive SP+ for three straight years, the Nittany Lions still allowed 27 or more points in one-third of their games.
"You've also gotta get lucky on a guy," Franklin said. "Not only do the guys have to pan out that you think [will], but a couple of 'em have to be better than you even thought. Obviously you saw what happened at LSU last year with Joe Burrow. Would anybody have anticipated that? And I don't mean that as a knock. I have so much respect for him and for LSU, but look at New England too. Bill Belichick is maybe the greatest coach in the history of the NFL, but Tom Brady, that was fortunate. No one thought that was how it was gonna play out.
"Whether you're able to just go get a no-brainer [recruit] that is a generational guy out of high school, and everybody can recognize he's a no-brainer, or you're fortunate enough to get lucky on a guy, either way you need enough of those difference-makers."
Granted, PSU quarterback Sean Clifford was more touted as a recruit than Burrow, and he has a couple of more years to reach his ceiling. But it bears mentioning they almost had a "generational guy out of high school" -- quarterback Justin Fields -- before he decommitted from Penn State, spending a year at Georgia, then ended up at Penn State's biggest current rival, Ohio State. You've got to get lucky and stay lucky, apparently.
In the absence of a generational QB, though, what can you do to keep pecking away at the ceiling your program seems to have? Math, of course.
"Before the 2002 season, Paul DePodesta had reduced the coming six months to a math problem. He judged how many wins it would take to make the playoffs: 95. He then calculated how many more runs the Oakland A's would need to score than they allowed to win 95 games: 135. Then, using the A's players' past performance as a guide, he made reasoned arguments about how many runs they would actually score and allow." -- Michael Lewis, "Moneyball"
The football analytics revolution is finally upon us. We have player tracking data at the NFL level. Like every other professional sport did years ago, every team is hiring an analytics director and scooping up Twitter nerds. At the college level, companies like Pro Football Focus are giving coaches access to tendency data and reports, and companies like Championship Analytics are providing decision-making guidance for fourth downs and other situations.
As teams get further down the analytics rabbit hole, now seems like a pretty good time to ask ourselves a really simple, really complicated question: How do you win football games? How do you find some spare points between the couch cushions like DePodesta and the long-ago Oakland A's did? Pondering that question, I decided to go old-school: I created a rubric.
I have for years leaned on a simple efficiency measure called success rate. If you gain 50% of your necessary yardage on first down, 70% on second down or 100% on third or fourth, the play is a success.
You could also call the play a win -- a successful play is a win for the offense, an unsuccessful play a win for the defense. Working through years of college football play-by-play data, I determined that every play you win on either side of the ball is worth about 0.3 points to your scoring margin. Plays in certain situations are worth a bit more. Turnovers, special teams, big plays and all the other things that a team can live or die by? They're worth something too. You can deconstruct a football game into these different events and figure out where you're leaking points and where you're doing particularly well.
Wins (plays): 0.3 points
Penalty: -0.4 points
Third-and-short wins: 0.6 extra points
Third-and-medium wins: 0.9
Third-and-long wins: 1.0
Fourth-down wins: 1.7
Red zone wins: 0.3
Punt attempts: -1.0 points (since you're officially forfeiting points on a given possession)
Gains of 20+ yards: 1.3 extra points
Gains of 40+: 1.5
Zero-yard gains or losses: -0.2
Turnovers: 2.6 points
Fumbles committed: -0.6
Defensive return touchdowns: 2.5
Good special-teams factors
Successful punt (by my punting success rate definition): 0.5
Punts with 50+ net yardage: 0.5
Successful kickoffs (resulting in field position at 25 or lower): 0.6
Punts with 45+ net yardage: 0.3
Field goal made: 0.4
Two-point conversion made: 0.9
Bad special-teams factors
Punts with under 10 net yardage: -2.0
Kickoffs with under 10 net yardage: -1.8
Field goal missed: -1.9
PAT missed: -0.7
In playing with both NFL and high school data, I found that you end up with pretty similar values. At the pro level, big plays, red zone wins and bad punts/kickoffs were generally worth less, and third and fourth downs were worth more. At the high school level, big plays and bad special-teams events had more value, and I found an extra value was important: wins inside your own 30. As with a sport like soccer, a lot of successful high school football comes from disaster avoidance.
Here's what this method produced for recent Penn State seasons:
2017: +24.6 PPG estimated scoring margin using the values above, +25.0 actual
2018: +15.4 PPG estimated, +13.2 actual
2019: +16.9 PPG estimated, +19.8 actual
Average: +19.0 PPG estimated, +19.3 actual
Other recent notables:
2019 LSU: +26.5 estimated, +27.7 actual
2019 Clemson: +33.2 estimated, +30.4 actual
2018 Clemson: +29.5 estimated, +31.1 actual
2018 Alabama: +27.2 estimated, +27.5 actual
2017 Alabama: +25.8 estimated, +25.1 actual
2017 Georgia: +17.8 estimated, +19.0 actual
Average of these six teams: +26.7 estimated, +26.8 actual
I'll write more about this system in the coming months -- I think it's got a lot of potential -- but for now, let's use these buckets to figure out where Penn State lacked last year, where it has made major gains in recent seasons and where it might be able to steal an extra point or two. To do so, we'll compare PSU's output to that of last year's top three teams in each group of factors.
2019: LSU +8.4 PPG, Clemson +7.4, Ohio State +7.0, Penn State +4.6
You want to know how much can separate the elites from the near-elites? On average, LSU, Clemson and Ohio State won 81.4 plays per game (offense and defense) to opponents' 56.0. Penn State's margin was 76.2 to 62.3. Winning just one extra play per quarter would have made it 80.2 to 58.3, almost completely catching the Nittany Lions up in this category.
One play every 15 minutes!
A specific type of play in particular held PSU back too:
2019: Ohio State +15.3 PPG, Clemson +11.4, LSU +11.0, Penn State +5.4
In the most valuable plays in a given game, PSU was consistently coming up a bit short.
On passing downs (second-and-8 or more, third- or fourth-and-5 or more), Clemson won an average of 9.4 more plays than its opponent in a given game, while Ohio State was at +8.2, LSU +8.1 ... and Penn State +1.7.
On third-and-long (7 or more yards to go), Ohio State was +5.1, Clemson +3.0, LSU +3.5 ... and Penn State +0.5.
On offense, the Nittany Lions were a decent 38th in standard downs success rate but 107th on passing downs. First-year starting quarterback Sean Clifford played like a first-year QB in these situations, and his receiving corps wasn't deep with options. KJ Hamler and tight end Pat Freiermuth combined for 99 catches, but only one other player (Jahan Dotson) had more than 15.
Even Penn State's defense struggled in this area. The Nittany Lions had maybe the best run defense in the country and ranked 10th in standard downs success rate allowed. But they were only 40th on passing downs, 62nd on third-and-long. Worse, they weren't merely allowing 7 yards on third-and-7 -- in third-and-long situations, Penn State allowed 23.6 yards per successful play, compared to the national average of 18.8. They were getting gashed.
How can you be so good overall but so shaky in one particular scenario? Sometimes it takes until the offseason to find the answer.
"We've been a tremendous sack team for five years now," defensive coordinator Brent Pry said. "And what happened in a few different games last year was, we were more spread out on those blitz situations -- we're not blitzing, we're rushing four, and it's more identifiable what the coverage is. What we found was, we were just too vanilla. Too easy, too identifiable for good playcallers.
"So it was identifying what we felt was the biggest issue and then coming up with some answers. We took the four or five best teams against empty and studied them. Last year it was defending the run game [that needed work], and we turned around and were No. 1 in yards per carry."
2019: Clemson +7.3 PPG, Ohio State +6.7, LSU +5.0, Penn State +2.5
Once Journey Brown emerged late in the season -- the then-junior rushed for 119 yards per game and 7.6 yards per carry over the last four games -- Penn State was just about as explosive as ever. Hamler averaged 16.1 yards per catch, Dotson 18.1. But the best offenses are explosive and consistent, and PSU struggled in that regard. The combination of offensive inefficiency and late-down defensive glitches meant that while Ohio State was averaging 7.0 gains of 20+ yards per game and allowing just 3.1, PSU was averaging 4.8 and allowing 4.2.
"Can you throw the ball when everyone in the stadium knows you have to throw the ball?" Franklin said. "When it's second-and-long, third-and-long, you've gotta be able to do it. And that's some of the stuff that I think we can script in practice a little bit more."
Franklin also made a very intriguing new offensive coordinator hire.
With Ricky Rahne taking over as Old Dominion head coach, Franklin brought in one of the OCs who had enjoyed some of the aforementioned passing success against PSU: Minnesota's Kirk Ciarrocca. The Golden Gophers ran a physical, run-heavy offense overall, but when opposing linebackers and safeties were wrong-footed, quarterback Tanner Morgan and receivers Tyler Johnson and Rashod Bateman torched them with slants and go routes. According to Sports Info Solutions, Minnesota had the 13th-highest percentage of RPO passes (as a percentage of all passes) and was devastating with them: 65% success rate (eighth), 23% explosive play rate (fifth). Better yet, the Gophers could pass when they had to. They were 10th in passing downs success rate. They caught up to the chains in ways that PSU could not.
PSU has the pieces of what should still be a dynamite run game, even without Brown. Noah Cain is ultra-efficient, Devyn Ford is explosive and the line should be as sturdy and experienced as any Franklin has had. But while Ciarrocca inherits one of the best tight end rooms in the country with Freiermuth, Zack Kuntz and exciting youngsters, the receiver position is ultra green without Hamler. The solutions he attempts in 2020 won't be identical to last year's in Minnesota.
"It's always personnel-driven," Ciarrocca said. "As you're game-planning, you know what your team's strengths and weaknesses are, and you're looking at your opponent's strengths and weaknesses and how you can exploit them.
"It's not broke here; it's not a rebuild. When you come into a situation like this, you're just looking for, how can we help the players perform just a little bit better in all the situations that come up -- first down, second down, red zone, coming out backed up, they all add up. You're talking about one point or one play a game that can swing it for you, and you just want your players to constantly improve and get better to make that one play that can swing your whole season. It could be the difference between making the playoffs and not."
2019: Clemson +4.0 PPG, Ohio State +2.5, Penn State +2.3, LSU +1.2
Last season, PSU was 18th in turnover margin and eighth in expected turnover margin -- which projects what the margin would have been with normal fumble recovery rates and interception-to-pass-breakup ratios. That put them in the ballpark of the contenders, and since passing downs have a bit of a correlation with turnover chances (due to both pressured passes and sack-and-strip fumbles), improvement in that regard should take care of the rest of the gap.
2019: Clemson +3.2 PPG, Ohio State +2.5, LSU +2.2, Penn State +2.0
Special teams is ripe for stealing a few extra points, and there are two ways to look at this category: (1) PSU still has a bit of ground to make up, and (2) the Nittany Lions have come an awfully long way. Their per-game average was -1.6 per game in Franklin's first three years but has been at +2.0 or higher each year since.
Last year, Franklin hired Memphis' Joe Lorig as his special-teams coordinator. They had coached together briefly at Idaho State 20 years earlier, and in recent years Lorig's success as a special-teams coach has given him a reputation. Memphis ranked first in special-teams SP+ in 2016 and averaged +2.4 points per game from special-teams factors from 2016 to '18.
"My job is as much sales as anything, in my opinion," Lorig said. "Building a culture and selling to the guys why [special teams] is important. Of course it's important! But how do I get them to really believe it? Another part is, I have to have the tools if you're gonna hold me accountable. So how many scholarships are we gonna allocate [for specialists]? Some schools allocate one, and there's a school I know of that allocates seven!" Having a full scholarship to give to a high school punter or long-snapper can make an obvious and enormous difference.
"Also, every guy on our team except for the offensive line and quarterbacks can start on two special teams [units]," he said. "How do you convince a star linebacker to also do a great job being the left guard on punt? How do you convince a star receiver or running back that, hey, you're also gonna be a great off returner? You may not have the ball in your hands; you're gonna have to lead up and lead block.
"Building that culture from the top down is very, very critical. Culture's a word that can be overused, but there absolutely is a culture. To do it right is really hard."
"In a given year in the NFL ... sixty to sixty-five percent [of teams] will cluster between two games above or under .500. As a result, it has been hypothesized statistically that as few as 6-10 plays a year will separate a team from finishing one game over .500 to one game under 8-8." -- Bill Walsh, "Finding the Winning Edge"
Parity does not exist in college football like it does in the NFL. The gulf between college football's ruling class and everyone else can sometimes seem enormous. It especially has early in the 2020 season, where the top teams have mostly looked the part and so many potential second-tier teams -- Oklahoma, LSU, Auburn, Texas, Florida -- have struggled.
For nearly elite programs like Penn State, Wisconsin, Michigan, Washington, Notre Dame and others, though, the gap might not be as wide as it seems, and it might not take a Deshaun Watson or Joe Burrow (or Justin Fields) to close it. One play per quarter ... an extra third-and-long stop ... one more 20-yard gain ... that might be all it takes.
Week 8 playlist
The slate in Week 8 is fuller than it has been all year! Here are 10 games -- at least one from each time slot -- you should pay attention to if you want to get the absolute most out of the weekend, from both information and entertainment perspectives.
All times Eastern.
Auburn at Ole Miss (noon, SEC Network): Auburn's Bo Nix lost the plot last week but faces an antidote in the Ole Miss defense; meanwhile, the Rebels face an even better D than the Arkansas one that frustrated them last week.
No. 17 Iowa State at No. 6 Oklahoma State (3:30 p.m., Fox): We keep saying that the Big 12 has all but eliminated itself from CFP consideration, but until OSU loses, that's not actually true. Can the Pokes keep it up?
No. 18 Michigan at No. 21 Minnesota (7:30 p.m., ABC): The narrative potential for this one is off the charts.
Texas State at No. 12 BYU (10:15 p.m., ESPN): Late-Saturday football again! BYU rediscovered its top form late against Houston and should roll, but watching the physical Cougars dominate is awfully fun.