We’re deep into spring practice now. The question is: Can you really learn anything of value from spring practice?
Yes. I think. Probably. You can certainly get a look at individual players. If a guy looks bigger or faster, that’s relevant. You can learn information about the depth chart. You might even learn something about tweaks to the playbook or scheme (or maybe not – what coaches do in the spring is not necessarily what they do on 3rd and 4 against their rival). But what I’m convinced you cannot do is make predictions about team success for the following season.
Some of the reasons for this are obvious and well-known. Practice is not a game; it doesn’t have the same level of pressure, for one thing. In addition, in practice you’re playing against your friends and teammates, and they know you and your plays (maybe even your play calls and audibles) very well, much better than any opponent. And what conclusions can you draw from a big performance by one side of the ball? If the quarterback goes 10-for-10 for 200 yards, that might be a good sign for the offense. It might be an equally bad sign for the defense. Or it might not mean as much as you think for either side of the ball, both for the reasons discussed above, and for another reason: small sample sizes.
A few scrimmage situations provide very limited information. Of course, football is a game of small sample sizes. MLB teams play 162 games. Strange statistical extremes tend to get ironed out over the course of 162 games. Baseball fans realize this. If Mike Trout hits .245 in the first twelve games of the season, nobody assumes Trout is washed up. Angels fans don’t start looking up the stats for the AAA affiliate to see if the centerfielder there is a better option.
If a news organization polled twelve people on an upcoming presidential election and got 9 votes for Biden and 3 votes for Trump, no reasonable person would assume that the popular vote will be in the neighborhood of 75-25 in favor of Biden. Almost everybody understands that small sample sizes mean predicting from the available data is very difficult.
But people also have a need to see patterns and find causation where randomness may very well exist. And therefore, while football is a sport with very small sample sizes, we like to pretend that we know more than we do.
We know that one of the best stats for judging the quality of a football team is yards-per-play differential, in part because there are lots of plays in a football season, and thus the stat can iron out some of the extreme results you find in small sample sizes.
Look at this chart from Bill Connelly, one of the sharpest football writers around:
Even small increases in yard per play margin gives a team a considerable edge.
Now I want you to compare the numbers from two seasons: last year’s 5-1 USC team and the 5-7 team from 2018:
You look at that chart and say, “Wow. There wasn’t a huge difference between 2018 and 2020. The 2020 team was only slightly better on that key metric.”
And you’d be wrong. The better numbers were from 2018. Last year’s team was downright mediocre in yards per play with a differential of only 0.1.
The 2020 team had a substantially better winning percentage but was downright poor on what many advance metrics people believe is a key statistic. So what explains the difference in the records? Largely one thing: turnovers. Last year’s team had a turnover margin per game of +0.8, which is good, and the 2018 team had a turnover margin per game of -0.8, which is terrible.
And none of this had anything to do with having a more-experienced QB or a better offensive scheme or anything else on that side of the ball. Because the 2020 team actually turned the ball over more, 1.8 times per game versus 1.6 times per game in 2018.
No, the entire difference in turnover differential – and, in some sense, the entire difference between a disastrous record in 2018 and a good record in 2020 – was takeaways per game. USC was second in the country last year at 2.7 per game and was 125th in the country at 0.8 per game in 2018. That’s the primary statistical difference between the two seasons. Turnovers can be the difference between 5-7 and 5-1. Over the course of six games, even twelve games, it’s not unusual to have extreme statistical disparities like this, just as it wouldn’t be unusual to have extreme disparities in presidential election polls if polling companies asked only six or twelve potential voters for their opinions.
Last year, USC won three games they could easily have lost. That’s the difference between 5-1 and 2-3, and without an appearance in the Conference Championship Game. In 2018, USC lost a series of close games. A bounce here or there and that 2018 team could easily have been 7-5. Based solely on the numbers, it would not have been crazy for the 2020 team to be .500 and the 2018 team to be slightly over .500. That’s what happens when you have small sample sizes.
With such a small sample size, a couple of plays of any kind can be the difference. If USC doesn’t get a series of breaks at the end of the ASU game, that game is a loss, and the entire complexion of the season changes. In a 162-game season, that just isn’t the case. One loss in the regular season means very little.
Now, I can hear some of you grumbling already. If it’s so hard to predict football because of small sample sizes, how do I know that Alabama will be good again? Simple: because the better the program, the fewer competitive games it will play. Alabama under Nick Saban and USC under Pete Carroll were substantially better than almost every team on the schedule. As a result, those teams would play relatively few competitive games – usually two or three a year. And they would lose about half of those games over the long term. If you play two or three competitive games a year, you will sometimes go 12-0, you will sometimes go 9-3, but you will more likely come in at 11-1 or 10-2.
If you’re an above-average program, like USC is now, you will play closer to six competitive games every year. With six competitive games, you can have wild fluctuations year to year. You probably should go around 9-3 or maybe 8-4. Over the course of 162 games, you probably would win the “correct” percentage of games as the disparities even out and the teams sort themselves out according to their talent and skill. But because of the small sample size, you might get more lucky bounces than usual and go 10-2, maybe even 11-1. Or you might get more bad bounces than usual and go 6-6.
And because the difference between the two is often (not always, but often) turnovers, which way the ball will bounce is difficult to predict. Some players and teams turn the ball over more than others, of course. Deion Sanders will get more interceptions than the average cornerback. Aaron Rodgers will throw far fewer picks. But the strange bounce of an oblong ball over the course of very short seasons means that you can expect fluctuations in turnovers, and it may not be because your scheme, practice intensity, or even players are substantially better or worse year to year. These are Alabama’s rankings in takeaways over the last four years:
1.6 (54th) in 2017
1.4 (79th) in 2018
1.9 (13th) in 2019
1.7 (43rd) in 2020
Alabama had a good scheme, good coaching staff, and good players every year. But it’s hard to control for turnovers over a short season. The difference between 13th in the nation and 79th in the nation was half a turnover a game – six plays in a season. A few bounces here and there can skew everything, including wins and losses.
What does this all mean? Am I saying that spring practice is meaningless? No, it’s a critical developmental time, and therefore spring practices are important to the players. We can even learn some useful things. Michael Jackson III is probably going to be a good receiver, for example.
But we can’t predict much about the final record in 2020. The fundamentals have not changed much at USC over the last few years. The team has more talent than most of its opponents, but not as much as the elite programs in the country. It’s a program that can win almost any game on the schedule (since there’s no Alabama, Clemson or the like on the schedule this year), but one that can lose to all but a couple of teams on the schedule as well. Like I said: above average, maybe even pretty good. Over the long term, I would expect USC to win about half of its competitive games and settle in for a winning percentage just under .700. But the 2021 season is not “the long term.” A couple of turnovers here, a blocked kick there, and a receiver getting his foot barely in-bounds instead of barely out-of-bounds can be the difference between 7-5 and 10-2. And we’re no closer to predicting which is more likely now than before spring practice started.
I don’t want to get into the details of the LA Times expose on the Song Girls. I’ll just say that I hope the facts come out and people are treated justly – whatever that means here – and I wish these young women the best with the difficulties they are going through.
But one big-picture question that comes to mind is: What are the Song Girls for? If it’s to be ogled by middle-aged men, that’s kind of gross. If it’s to demonstrate amazing dance steps and skills, that’s kind of weird. USC football games are not performances of the New York City Ballet, nor are they dance recitals. It can’t be to lead cheers. The Song Girls don’t lead cheers, and the days are long since past when most crowds at most sporting events engage in much disciplined, directed cheering anyway. So why Song Girls?
Cheerleaders from the beginning have been associated with physical attractiveness, and there are probably a whole host of cultural reasons for that, all of them beyond this random thought. The Song Girls probably can’t escape that cultural expectation. I suspect most Song Girls don’t try to. But there’s more to it than that here. The Song Girls are not ordinary cheerleaders; they are iconic, I think because the Song Girls are a depiction of what USC likes to project to the world. They are Show Biz; they’re Hollywood.
And that means aesthetics have always been an important part of the package, just as aesthetics matter for actors, pop stars, professional dancers, and a great many other entertainers. Brad Pitt and Jennifer Aniston are actors, and they’re undoubtedly good at their craft, and there is no necessary connection between acting skill and cheekbones. But we know cheekbones matter. George Clooney is a good actor. But so is Danny DeVito, and he’s not getting George Clooney’s roles.
The Song Girls, like other entertainers, have to work hard to develop their skills, and what they do requires talent and training – and how they look is an important part of the job, and everybody knows it, whether we want to admit it or not.
That doesn’t tell us whether these young women were treated appropriately or not. That will come out in time, I suspect. And it doesn’t answer the normative questions. Should the Song Girls be judged on aesthetics? Is it appropriate for a university to put its female students in a position where they are being judged on their bone structure or their legs? Is it a good idea for people to take on roles where they subject themselves to scrutiny over their appearance, especially when it’s such a struggle for young women to have a healthy self-esteem when it comes to body and appearance issues? I’m not going to try to answer any of these questions. But I can certainly understand how this controversy would arise. I’m surprised it doesn’t come up all the time with a lot more cheer and dance squads. Or maybe it does, and I just don’t read the right publications.
With Erik and Greg doing an excellent job of covering USC’s spring practices, I thought I’d take some time to look into the Trojans’ conference foes. My plan is either to make it through the entire conference or get bored and stop early. I’m not sure yet.
First up: the Bay Area schools.
Record against USC: 33-62-3
Program motto: Intellectual
Defining program moment: Having a trombone player run over by a Cal Bear in the end zone
Stanford is one of the elite academic institutions in the world. As such, Stanford just doesn’t care about plebian pursuits like football. Well, on those rare occasions when they’re good, they care. But when they’re bad, they assure everybody that they actually don’t care at all, football being beneath these future titans of academia and industry. So I guess most of the time they don’t care.
Coach David Shaw is one of the most respected coaches in the conference, even though his program is not as good now as it was when he took over. I suppose we should give him credit for his ability to pump the breaks on inevitable mediocrity. When Jim [Harbaugh] left, everybody thought Stanford would immediately go back to being mediocre. But David Shaw pushed back, rode the fumes for a few years, and delayed Stanford’s fall. But, guess what? It certainly looks to me like they’ve fallen.
There’s an upside, I suppose. Early in his tenure at USC, Clay Helton famously talked about how USC football aspired to be like Stanford. (Excuse me. I just threw up in my mouth a little.) And it appears the Trojans have done just that. Over the last three years, the Cardinal are 17-14 while USC is 18-13. Cue Conquest?
Stanford is famous for its game-day atmosphere. Or maybe not. It generally has a half-empty stadium that is largely filled with visiting fans. But the university does take great pride in its distinctive marching band. This is one of life’s enduring mysteries. The band dresses badly, sounds worse, and routinely fails to meet what are already shockingly low expectations for class and musicianship. Maybe they were the nation’s first punk marching band? Maybe they were (and are) entitled kids who don’t put in the time or effort to be any good. Maybe I don’t care and I just want them to go away. Yes, I think that’s it.
Future projection: A football team that increasingly plays like the band looks. I see fourth or fifth place finishes as far as the eye can see.
Cal Golden Bears
Record against USC: 30-71-5
Program motto: Workers of the World, Unite?
Defining program moment: Running over the Stanford trombone player
Cal is also an elite academic institution. That’s good; Its academic standing and reputation are certainly no reason for shame or embarrassment. Yet I get the sense that there is a certain amount of shame and embarrassment at Cal. The problem is simple: it’s not Stanford. Being Jermaine Jackson is cool, I suppose. But it’s less cool when your brother is Michael. Cal and Stanford are rivals, and Cal students want to think they attend an institution that is Stanford’s peer. But they don’t, not really. The whole world knows it. Deep down, even Cal students know it. And that’s probably why they are so angry all the time and throw batteries from the stands.
Cal is coached by Justin Wilcox, and I’m not sure what to think of him. Wilcox’s defenses at USC were so bad that even Clay Helton found them intolerable. The defense wasn’t disciplined and didn’t tackle, and that was apparently Wilcox’s fault. USC fans certainly weren’t clamoring to keep Justin “Young Monte” Wilcox on the staff. I would have driven him to the Greyhound station myself if asked. (That offer is still available if anybody can think of coaches who need and deserve such help.) Yet Wilcox has fielded some very good defenses at Cal, to go along with some very bad (I mean, really bad) offenses. What this means is if you can score twice – and you probably can; their defense isn’t that good – you’re probably safe from an upset.
Aaron Rodgers played at Cal, of course. Aaron Rodgers is weird. Insanely talented, kind of fun, and really weird. I don’t know if Rodgers was weird before he attended Cal or if attending Cal made him weird. I’m hoping to get that information to you before I lose interest in this series.
Future projection: If Aaron Rodgers returns, Cal might do something. Because I don’t think he can or even wants to, they won’t. Cal will continue to be Oregon State with better SAT scores and less-awful uniforms.
Carthago delenda est.