Bumped: Another thoughtful look from our community at some stats and how they might predict 2015. - BN eds.
This morning, Bill C. over at Football Study Hall released his final preseason S&P+ rankings in preparation for the season, which you can find at this link. The point of this fanpost is to break these numbers down to better set some season expectations.
First, let's start with some precautions. Advanced stats for football are much less predictive than they are for other sports such as basketball and especially baseball. The reasons are vast: there are 22 individuals on the field at any one time, the small sample size of playing so few games, the odd shape of the ball leads to randomness (especially with fumbles), etc. That being said, these stats are rather good at setting a baseline idea for how a season should go, and especially future seasons. In Bill's excellent book Football Study Hall, Bill points out that in 2011 both Notre Dame (8-5 that year) and Texas A&M (7-6) were ranked in the top 15 of the F/+ rankings. For both teams, their record showed disappointment, but the underlying story was that both teams were on the verge of breaking out, which they did in 2012 (Notre Dame ended up in the BCS Title Game, while A&M went 11-2 and beat Alabama on the road). The point being, while these numbers are not the be-all, end-all, they can tell a different story than a record might indicate.
So, with that being said, let's start by looking at how S&P+ numbers are formed. From the article:
The process is to come up with three sets of projections based on five-year performance (weighted to make 2014 more important than 2013, and so on), the likely changes associated with each team's returning starter figures, and the likely changes associated with each team's two-year recruiting averages (recruiting rankings are relatively solid predictive stats).
I blend them together based on what has produced the best results in the past. That means recent history carrying the most weight and recruiting carrying the least.
S&P+ is designed to put more weight on recent success rather than past success, but does still take into account past history, which is indicative of trends and possible future outputs. Returning starter figures makes sense as an obvious component, as those starters will have experience, and the better those starters were, the better you can make them out to be in the future. And, as UCLA_beer&mathmatics pointed out last week, while talent level isn't indicative to success, having elite talent is usually a crucial component to a winning formula (as an aside, Grantland's Matt Hinton pointed out in his article from last week the same basic idea, noting that every NC winner since 2012 ranks in the top 10 of recruiting rankings. The point is simple: yes, coaching can make a difference, but especially at the college level where talent is distributed unequally throughout the NCAA, the teams with better talent will outperform other teams on average).
With that being said, let's break down the numbers on both sides of the ball before looking at the overall story. First, let's start with the offense.
Depending on how you feel about Noel Mazzone and his playcalling ability, Adrian Klemm's OL-coaching ability, or the fact that we're starting a true freshman, the preseason projections may surprise you. UCLA is ranked 5th overall in offense for the coming season. There are a few reasons this is the case:
1) UCLA's offense, statistically, has been very good the last few years - This is really one of those "Stats don't tell the whole story" cases, but the offense has put up very good numbers the last few years. UCLA's offense finished 8th in S&P+ last year, following performances of 18th in 2013 and 26th in 2012. We know the offense has laid some eggs in the past few years; the performances against Stanford and Utah in particular come to mind, as does the first half against Oregon. But, these aren't outside the norm, and going against expected offensive output, the offense succeeds more often than it fails. And S&P+ isn't the only formula that loves the Bruins' offense; Brian Fremeau's FEI (Fremeau Efficiency Index) also loves UCLA, ranking the offense 6th in 2014, ahead of teams like Ohio State, Florida State, and Baylor.
2) There is a lot of returning talent - This shouldn't come as a surprise. UCLA returns 9 linemen from last year's 2-deep, 5 of their top 6 receivers (including the top 3 leaders), and, oh yeah, the running back that led the Pac-12 in rushing yards last year. Even with an unproven quarterback, that much returning talent is hard to overlook, to the point where Josh Rosen might not be asked to do much in his first few games while he gets his feet underneath him.
3) The incoming talent the past few years is surprisingly good - Much has been made of UCLA's seeming inability to recruit elite offensive talent, specifically a true game-changing slot receiver. That said, the offensive talent the past few years has been maybe a tad underrated. Last year's recruiting haul had, of course, the nation's #1 QB, TE, RB/ATH, and Offensive Line class, but look at 2014's recruits. Nate Starks was a revelation at RB, providing a great compliment to Perkin's ability. Mossi Johnson was a breakout star as a true freshman last year as well. Meanwhile, guys like Austin Roberts and Jordan Lasley got the chance to redshirt and bulk up, and suddenly UCLA's skill position's look even better than they did at first glance.
4) A full year of McDermott - It really cannot be overstated just how much Connor McDermott changed UCLA's offensive line last year. Through the first six games, with McDermott out of the starting lineup, UCLA gave up 25 sacks. Over the final 7 games with McDermott in the lineup, UCLA only gave up 15. That's a per-game average drop from 4.17 to 2.14. This next part is a bit of projecting, but with Rosen being much more of a pocket passer than Hundley and more willing to throw the ball out of bounds than hold it in an attempt to make a play, but I'm cautiously optimistic that this improvement will continue into 2015.
Remember up top when I pointed out that previous results were taken into account in the S&P+ rankings? Well, that's very apparent with the defense. UCLA's defense ranks 21st overall in S&P+, and again, there are a few reasons at play here:
1) Ulbrich was a fine coordinator...but the talent level was too high for fine - Let's be fair to Ulbrich first: UCLA's defense ranked 25th last year in S&P+, only down one spot from 24th in 2013. There were two absolute stinkers defensively (Oregon and Stanford) and a few absolutely wonderful performances (against Virginia, Arizona, and USC). That said, the biggest indictment against Ulbrich would be the defensive performances against Memphis, California, and Colorado. Against those three teams, UCLA had a very clear and obvious talent advantage, but Ulbrich's passive defense allowed those teams to stay in games and forced the offense to operate at a high level (to be fair here, Memphis did end up as a really good team, but for an early season home game, that defensive performance was still unacceptable). The defensive front-seven really didn't create any pressure, and the prevailing thought was that Ulbrich was counting on talent and gap integrity to win those battles. That's a fine strategy in the NFL, but in college you have to be more proactive in creating mismatches and pressure, and Ulbrich was burned for it more than a few times.
2) Bradley is an upgrade, but isn't the one playing - This is an unfair criticism to be sure, but the fact is that while Coach Scrap is a huge upgrade from a coordinator position, he's also not the one on the field making the plays. Which means that, for the purposes of compiling S&P+, the coordinator change had a negligible boost to UCLA's ranking. This same paragraph also applies to the change to Scott White at LB coach, which is also a clear upgrade, especially from a recruiting standpoint.
3) Talent is there, but needs to break out - UCLA's defensive recruiting has exceeded the offensive recruiting under Mora, which is understandable considering Mora is a former defensive coach. That said, while there has been some breakout talents on defense (hello Myles Jack) the unit as a whole has underperformed relative to the talent level. This is especially true on the defensive line, which really had issues creating any sustained pressure throughout the season and was average against the run (that UCLA's pass defense ranks as well as it does despite the lack of pressure is a testament to the talent and skill in the secondary).
Combined, the Bruins end up ranked 6th overall in S&P+ heading into the 2015 season. That puts them second in the Pac 12 behind Oregon (4th) and ahead of preseason South favorite USC (12th). What can we glean overall from these numbers heading into the season?
1) The defense will be better than their preseason rankings - Again, because S&P+ factors in previous seasons, UCLA's defensive numbers are only in above-average territory to start the season. Normally, when a team loses a defensive coordinator producing a top 25 defense, you expect a backslide, but its also not often you can say losing a coordinator works to the team's benefit as much as Ulbrich's departure did. Tom Bradley has a track record of producing highly-ranked defenses, and I'd expect UCLA's to be at least top 15 by season's end.
2) Even with a true freshman QB, UCLA's offense should perform....or else - Talent-wise, even the presence of a true freshman quarterback shouldnt hold the defense back too much. There's just too much talent returning. That said, Noel Mazzone is going to have to have a season to prove his doubters wrong. Despite how much Brett Hundley meant to UCLA Football, there has always been scattered reports of having to adjust game plans for Brett's skillset and abilities. I'm not landing on either side of the debate here, but with a QB that is more geared to Mazzone's playstyle, he's going to have to prove that he can keep the offense running at a high level. The stats back up the idea that this can be the case, but he's going to have to prove it on the field.
3) The BYU game should be the bellwether game for the season - We're not going to learn much from the Virginia and UNLV games to start the season, but they should be fine tuneups and help get the team, and especially Rosen, into an early groove. BYU, however, poses a significant threat. Defensively, they'll be one of the top defenses UCLA will play all year, especially against the run, and offensively Taysum Hill is experienced and deadly, so the Bruin defense will need to step up and prove they can stop a running quarterback (a problem in recent years). UCLA should win this, but the how is going to be rather telling for the rest of the season.
4) The first half is tough - UCLA's front-half of the season is loaded. Past the first two games, UCLA plays a tough game against BYU, then follows it up @ Arizona (S&P+ Rank 36), home against ASU (23), and @ Stanford (18) on a Thursday. If UCLA can get through that stretch undefeated, not only would they have climbed an important program bench mark (actually beating Stanford) but they'll have set themselves up for a special run. The second half of the season has a run of 4 opponents UCLA should beat easily before going to Utah and USC, but the first half is going to decide whether this is a good season or a truly special one.
5) UCLA's floor is much higher than it used to be - It's been a long time since you've been able to look at a season at UCLA and say that anything less than 9 wins would be considered a disappointment, but that's a testament to the job Mora has done at UCLA. With the talent on hand, 8 wins should be in the bag; 10 wins is easily attainable, and undefeated going into a conference title game isn't out of the question. But, that also means expectations are higher, and while Mora has proven he can bring in recruits at an elite level, he's going to have to start winning big games to silence his critics. There's no better year to do that than this year.