Analyzing the 2015 College Football Season using Big Play Percentage and Toxic Differential

WHICH TEAMS WERE THE MOST EFFECTIVE AT GENERATING AND PREVENTING BIG PLAYS LAST YEAR?

As a college football fan, you don’t have to read too far into Danny Kelly’s Ringer article on advanced statistics and the NFL before you can’t help but wonder how some of the stats mentioned would apply to the amateur version of America’s most popular sport. A quick Google search for NCAA football advanced stats proved fruitless, and ProFootballFocus.com and FootballOutsiders.com, the two main sites referenced in the article, are both primarily focused on the NFL and don’t yet offer comprehensive NCAA data. The realization that not many people out there are tracking this kind of stuff for the NCAA led to a second question: Why? The differences in play style are even more exaggerated in college than they are in the pros; you’d think that would make percentage stats that try to account for those differences even more popular among teams and fans, but it’s actually almost impossible to find anything outside of your standard counting stats, and even those leave something to be desired.

Of the 5 stats mentioned in Kelly’s article (Big Play Percentage, Toxic Differential, Accuracy Percentage, Running Back Success Rate, and Pass Rush Productivity), only 2 (BP% and Toxic Differential) can be calculated using resources readily available online.

Accuracy Percentage, which calculates a QB’s true passing accuracy by excluding spikes and throwaways and counting drops as catches would be extremely informative in college where drops are exceedingly prevalent, but you can’t even find comprehensive stats on dropped passes, let alone spikes/throwaways.

Running Back Success Rate adds context when analyzing running games by dividing successful running plays by total running plays, and is arguably a better metric than yards per carry. RBSR defines a successful play as gaining 40% of the yards to go on 1st down, 60% on 2nd down, and 100% on 3rd down, so even though that 2-yard carry on 3rd and 1 would lower your yards per carry, it would still be considered a successful play. This would be useful in analyzing the effectiveness of both individuals and teams, but unfortunately it’s either not tracked or not found easily anywhere on the internet.

Pass Rush Productivity does a better job of capturing pass rush disruptive than sacks alone because it also factors in things like QB hurries and knockdowns. This is one that I know coaches track internally, but with nowhere online reliably tracking knockdowns you can’t get all the way there without charting everything yourself.

That leaves us with Big Play Percentage and Toxic Differential, which both look at how effective a team is at both generating and preventing big plays. I think we can all agree that in general, both of these concepts are crucial to being a good football team, but I wanted to use these stats to analyze the 2015 season so that we can see who did it the best.

Big Play Percentage measures how effective an offense is at being aggressive by counting the total number of big plays and dividing by total plays. According to the article, a Big Play is defined as a run of 10+ yards or a pass of 25+ yards, and some teams have found that if you can just hit one of those in a given drive you have up to a 75% chance to come away with points. This makes sense logically; it’s much easier to score if you can gain yards in big chunks than it is to try and nickel-and-dime because it reduces the amount of opportunities for things to go wrong. Toxic Differential expands on this idea and applies it to both sides of the ball, adding Big Play Differential (Big Plays For minus Big Plays Against) to Turnover Margin to find out how effective teams are overall at generating and preventing big plays.

I used data captured on CFBstats.com to calculate both metrics. Due to the nature of the data they provide, I adjusted Big Play Percentage slightly to include passes of 20+ yards as big plays (instead of 25+), but otherwise left everything else the same. Here’s the spreadsheet:

Big Play Percentage and Toxic Differential Raw Data

First, some quick national observations:

As is tradition, the Big 12 sported the nation’s most explosive offenses with a League Big Play Percentage of 13.52%, while the Big 10 had the least explosive offenses in the country with an average of just 11.36%.

When it comes to stopping the big play, it probably comes as no surprise that the best defenses reside in the SEC, where teams gave up explosive plays only 10.79% of the time. On the other end of the spectrum, Big 12 defenses surrendered big plays 12.65% of the time.

The SEC had the best average Toxic Differential by an almost 2:1 margin, which lends credence to the idea that they are the nation’s best conference. In fact, the P5 rolls out as follows:

  1. SEC (18.21)
  2. ACC (11.36)
  3. Big 12 (11.20)
  4. Pac 12 (10.92)
  5. Big 10 (5.71)

That’s probably pretty close to how you would’ve ranked them last year, right? Also, while this isn’t any definitive proof that the Big 10 is the worst Power 5 conference, it does suggest that they’re by far the most boring. No big plays, no turnovers. Just 3 yards and a cloud of dust. Sounds fun.

BIG XII

Since I’m a West Virginia fan, I figured I might as well take a closer look at the Big 12, as well. The interactive table can be found here, but here’s a pic:

Big 12 BPP

Baylor had by far the most explosive offense in the conference last year and were actually the second most explosive unit in the country, with their 17.79% BPP second only to Oregon’s 18.93%. The majority of their 196 explosive plays came on the ground (131), but what’s crazy is that if you look at their month to month splits, they got the majority of those big running plays in 3 months. They averaged 12.4 explosive runs per game in September, October, and December, but only 4.75 per over 4 games in November, and if you think about it, this follows their season narrative: dominant early, uncertainty following Russell’s injury as they tried to figure out their new identity, then dominance again once they decided they were just going to grind teams into dust. What’s even crazier is that they return almost of all of that production on the ground. Baylor will have to figure out who’s going to step up and replace Corey Coleman and Jay Lee’s big play production in the passing game (they accounted for 55% of explosive receptions), but they do have KD Cannon back (15 big play catches) and should be able to lean on that ridiculous ground game while the new receivers find their feet.

Iowa State generated big plays 12.54% of the time, which is close to the national average rate of 12.21%, but had a tough time dealing with Big 12 offenses, allowing explosive plays at a rate 1.5% above the national average. The good news for the Cyclones is that they return the majority of the parties that were most responsible for their big plays. Mike Warren and Joel Lanning accounted for 75% of their big plays on the ground last year and should only get better, and the tandem of Lazard (14 explosive receptions) and Wesley (5) should improve on the outside, as well. The bad news is that it’s not going to get any easier to stop anybody. ISU’s Toxic Differential was poor, as well, and even though they return a nice amount of talent, a bit of improvement in the turnover department will be necessary for a more successful 2016.

Kansas had by far the least explosive offense in the league with a BPP of just 8.47%, and if not for Texas’ anemic passing attack, they would have had the lowest totals in both explosive runs and passes. Their defense wasn’t any better, surrendering big plays on 15.93% of snaps. The impact of this combination is made apparent in their -85 Toxic Differential, which was spared from being the worst in the country by UCF’s -93. Kansas had a bad-but-not-awful -7 TO margin, so we can deduce that they gave up 78 more big plays than they generated. The Jayhawks have some decent guys on offense who should improve, but idk how you find the bodies to make up for a disparity that big.

Kansas State was only slightly better than their in-state rivals on both offense and defense, hitting big plays on 10.84% of snaps and allowing them 13.93% of the time. Their -3 TO margin is uncharacteristically bad for a Bill Snyder team, but it’s the -39 Toxic Differential that most illustrates that disparity between their offensive and defensive BPP. The good news though is that they return a lot of production on both sides of the ball. Both Joe Hubener and Jesse Ertz are capable QBs, running backs Jones, Simon, and Dimel accounted for 31 explosive runs last year and should figure to be improved, and Burton and Heath should be better on the outside. If the Wildcats return to form in the turnover department and make it harder for teams on defense, they should be much better this year.

Oklahoma predictably performed well across the board last season, with their offensive BPP of 14.71% and defensive BPP of 10.11% good for 2nd and 1st in the conference, respectively. Their +58 Toxic Differential was very impressive as well, and good for 13th nationally. Overall the Sooners look in good shape going forward as well. Offensively, Mayfield, Perine, and Mixon accounted for 88% of their explosive running plays last year and were all top 20 in the league. They’ll need somebody besides Dede Westbrook (12 explosive catches) to step up and replace Sterling Shepard in the passing game, but considering their running game and Mayfield’s playmaking ability I think they should be fine. Defensively the Sooners return 2 great safeties and a good chunk of the front 7, but they did lose Zack Sanchez to the NFL so they may need some new playmakers to step up on the outside. All things considered, Oklahoma should again be the class of the conference.

Oklahoma State was actually fairy average in both offensive (13.41%) and defensive (12.05%) BPP when considering the season they had last year, which I guess is where the “I can’t believe people think Ok State is the conference dark horse” crowd comes from. The Cowboys also had a respectable but unimpressive Toxic Differential of 23, which like their defense was good for only 5th in the league. However, they do return a bunch of production on both sides of the ball. 81% of their big plays on the ground return, with the only notable loss being JW Walsh, and Washington and Ateman (18 and 11 explosive catches, respectively) are arguably the best 1-2 combo in the conference on the outside. If the Cowboys want to have a truly special season though they’ll need to find some balance, both offensively and as a team. The passing game should be fine, but they need to figure out how to inject some life into a running game that only managed 59 explosive plays last year (same number as Wendell Smallwood by himself), as well as figure out who’s going to fill Emmanuel Ogbah’s considerable shoes on D.

TCU was another one like Baylor where you wonder what could’ve been if they could’ve just stayed healthy. Attrition took its toll though, and by the end of the year they were bang average both offensively and defensively, with an offensive BPP of 13.45% and defensive BPP of 11.85%, good for 6th and 4th, respectively. A 4th place Toxic Differential of 26 would only serve to add to the mystery of how they were so good, until you remember that they had Trevone Boykin and Josh Doctson. Looking ahead to 2016, the Frogs lost by far the most production in the conference. Not only do they have to replace the aforementioned Boykin and Doctson (20 explosive catches), they also lost the wildly underrated Aaron Green (32 explosive runs) and Kolby Listenbee (13 explosive receptions). The biggest shoes to fill are unquestionably Boykin’s though. He did everything for them, and was not only the league’s most explosive passer in terms of big plays per game, but was its 12th most explosive runner, as well. TCU is returning a ton of talent on defense and Johnson and Hicks are both capable running backs, but it remains to be seen if Kenny Hill and those young receivers are capable of picking up all that Boykin/Doctson/Listenbee slack.

Texas had the least explosive passing offense in the league last year, but thanks to a strong ground game its 14.21% offensive BPP was actually 4th best in the conference. The problem was on defense. The 12.95% defensive BPP isn’t terrible in and of itself (not good either), but they gave up enough big plays (125) that even a +11 TO margin wasn’t enough to push their Toxic Differential into the positives. This was probably to be expected though with so many youngsters and first-year starters, and you’d expect that group to be much better this year. You have think their offense should improve, as well. Jarrod Heard (25 big runs), Tyrone Slopes (17), D’Onta Foreman (16), and Chris Warren (12) are all back to anchor the ground game, and if new QB Shane Buechele can provide a little balance through the air the Longhorns will be in much better shape offensively. They might still be a year away from being a real threat to contend again, but I expect them to be very competitive.

Texas Tech’s 158 explosive plays (91 run, 67 pass) were good for 2nd most in the conference, but unfortunately, and as I’m sure you can guess, the 153 big plays that they gave up were the 2nd most as well, and just 1 shy of Kansas’ 154. The biggest problem was the run defense, which allowed 106 runs of 10+ yards. I think about 13 of those came courtesy of WVU, so you’re welcome for that. Jokes aside, this is an issue that needs solving if Tech is going to be competitive this year. There are simply too many good running attacks in the top tier of the league, and you can’t break into that group if you can’t stop any of them. There are some question marks offensively, as well. Mahomes returns and figures to be a Boykin-like figure (1st in total big passes, 10th in runs), but gone are DeAndre Washington (49 explosive runs) and Jakeem Grant (18 explosive catches). Justin Stockton (13 explosive runs) will be asked to carry the load in the backfield, Davis (9 explosive catches), Lauderdale (8), and Sadler (7) return on the outside, and reinforcements are on the way in the form of a pair of high profile recruits, but it remains to be seen if any of them can provide the same sort of every down excellence and reliability that they’re losing in Grant and Washington. Overall, it looks like the status quo will be maintained in Lubbock; the offense will be good enough to win them a couple games they probably shouldn’t, and the defense will be bad enough to lose them a couple that they probably shouldn’t.

Finally, we get to West Virginia. Last but not least West Virginia. Best fucking Virginia. The Mountaineers had the second most explosive running attack in the league last year, but only the 6th most explosive passing attack, placing their overall offensive BPP of 13.89% right in the middle of the road. The defense was all-around excellent, allowing big plays on only 10.94% of snaps (3rd in Big 12) and leading the conference in takeaways with 31. This led to a very clean Toxic Differential of +49, good for 3rd best in the conference by a wide margin. For those wondering wtf happened for us to go 7-5, like with TCU, we can again look at the QB. It’s not the first time I’ve made it, but comparing our two rosters and performances from last year, I think you can make a strong case that the difference between Boykin and Howard was the difference between 10-2 and 7-5.

I’ll also use that Toxic Differential number to make the case that we should’ve finished 3rd in the conference last year. The Oklahoma State game at home (won the explosive play battle 7-4) and the Kansas State game on the road (9-4) stand out as two in particular where we played better than our opponents but failed to win because we made enough crucial mistakes that the other teams didn’t need to make big plays to win. I’m obviously bitter as hell and understand that shit happens, but you can’t argue that people would be taking our roster much more seriously if we win both of those and are coming off 10-3 instead of 8-5.

As it is, we’re looking at a lot of potential, but also a lot of uncertainty heading into this year. The running game was strong last year, but lost the biggest piece (Smallwood and his NCAA-leading 59 explosive runs). Rushel Shell had less explosive runs (17) than Skyler Howard (19) last year; will he be able to pick up the slack, or will we have to turn to Kennedy McKoy and Justin Crawford to give the running game some juice?

Conversely, the passing game was suspect but returns almost everyone. Shelton Gibson (14 explosive receptions), Jovon Durante (6), KaRaun White (5), and Gary Jennings can all make big plays, but can Skyler be consistent enough for our downfield passing attack to take the collective step forward from sometimes nuisance to full-time problem?

There’s even more uncertainy on the other side of the ball. Tony Gibson preaches takeaways and limiting big plays, so while he’ll be thrilled with our performance last year, I’m sure he’ll be worried about how we’re going to look with a unit that’s breaking in 7 new starters. We were the best team in the conference at limiting big running plays; can we replicate that performance with 3 new linebackers? Adding to those concerns, we return only 4 of our league-leading 23 interceptions on the back end; will the new group be able to produce at a level close to that if the front 6 isn’t consistently putting teams in bad situations again?

The good news in those first two cases is that we return 4 starters and 5 contributors from one of the best offensive lines in the conference last year, so if nothing else there should be room to run and time to throw. I’ll back our skill players to TCB in those situations. Defensively though, there’s less to feel warm and fuzzy about. I like our talent and the depth, but we desperately need some people to step up and make plays. Turnovers were our lifeblood last year, and it will be interesting to see how aggressive Gibby is able to be with all those new faces.

 

One comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s