- Basketball coaches have fully embraced the analytics era. What is it about college football that has made strategic innovation come in fits and starts? Plus, an early look at Oklahoma's closest Big 12 challengers and the rest of this week's #DearAndy mailbag.
One of our favorite high school football coaches was the subject of a discussion on Showtime’s Billions on Sunday. Kevin Kelley, the coach at Pulsaki Academy in Little Rock, Ark., is famous for rarely punting, always onside kicking and usually winning.
After watching Kelley’s Bruins snap Highland Park High’s 84-game win streak on a visit to Dallas three years ago, I wrote a column wondering why some college coach looking to close a talent gap and stave off an almost inevitable firing didn’t adopt some of Kelley’s methods. It appears I’m not the only one with this question…
From Matt: Are you surprised football coaches still don’t utilize some of Kevin Kelley’s methods despite mounds of mathematical data proving it works? NBA/college hoops has gone completely to the three-point game because data proved it works. Why won’t football coaches?
I am surprised. FBS coaches have embraced analytics to a point. Almost all of them subscribe to some sort of advanced stats service, and the largest staffs have analysts grinding through game video and crunching numbers so that the team might become incrementally more efficient. That’s really the goal of analytics no matter the field—to allow an organization to deploy its resources in the most efficient way possible.
Alabama’s Nick Saban has the largest staff, and the Crimson Tide typically are among the nation’s most efficient teams on offense and defense, but it’s difficult to discern how much of that is because of statistical analysis and how much is because Alabama usually has much better players than its opponent. The more obvious uses of analytics are a bit further down the talent food chain.
Chip Kelly turned Oregon from a pretty good Pac-10 program to one of the nation’s best programs by discerning through analytics that running plays at a higher tempo tended to wear out defenses. He then exploited a change in the clock rules to spring an offense on the sport that it clearly wasn’t ready for. At the same time, he ignored traditional stats in favor of ones that helped him better understand how that offense affected the entire team. He knew that because his offense ran more efficiently, his defense would spend more time on the field. So instead of worrying about total yards, he worried about yards per play allowed and points per drive allowed. This allowed Oregon to maximize its talent under Kelly.
But the difference between Kelly and Pulaski Academy’s Kelley (note the extra “e”) is that Kelly’s Oregon offense didn’t completely laugh in the face of conventional wisdom. It was more palatable to coaches, and that’s why so many of them copied concepts Oregon ran and added them to their own offenses and defenses. A closer comparison is the offense Baylor ran under Art Briles. (We’re just talking about on-field stuff here.) The Bears positioned receivers outside the numbers, stretching their formations across the entire width of the field. This took away the out route because outside receivers didn’t have room to run it, but it also forced most defenses to declare how many players they would devote to pass coverage and how many they would devote to run support. The quarterback could count the box and then choose to attack where the defense was weakest. But other than the coaches who worked in that system — Syracuse coach Dino Babers and Tulsa’s Philip Montgomery, for example — few were willing to go that wide because they considered it too extreme. One notable exception is Willie Taggart, who saved his job at South Florida in part by going that wide and then brought that offense to Oregon and now Florida State.
Matt mentioned that basketball coaches have embraced analytics wholeheartedly, and he’s correct. Look no further than the two teams that played for the national title last week. Coaches John Beilein and Jay Wright built Michigan and Villanova to take the most efficient shots (three-pointers and dunks/layups because midrange jumpers are statistically worth less and therefore are for suckers). They tend to play four guys who can shoot from the perimeter and take defenders off the dribble alongside a post player who can rebound and knock down an open three. This spreads the floor—spacing is key in both sports—and allows the Wolverines and Wildcats to have the easiest access to those ultra-efficient shots.
But there are more different ways to be efficient in football than in basketball. Plus, simply acquiring more talent creates a larger advantage in an 11-on-11 game than in a five-on-five game. But that doesn’t mean football coaches shouldn’t try to make the math work in their favor.
At Pulaski Academy, Kelley uses analytics a lot. For example, expected points helped him determine that it’s almost always better to go for fourth-and-short in the middle of the field. But there’s also an emotional component to his preferred style. By playing every down like his team is down 10 with 90 seconds to go, it forces a defense to adopt the mindset that it is protecting a 10-point lead with 90 seconds to go. How many times have you seen a defense give up huge plays (and points) in that exact situation? Part of what Kelley does is based on math, but part of it is designed strictly to tighten the sphincters of the people on the other sideline.
The style probably would have to be tweaked at the college level because college players are less apt than high-schoolers to make drive-killing mistakes, so when they got short fields by stopping the offense on fourth down at its own 30, they probably would convert that stop into points more efficiently. College teams also tend to have competent long snappers and punters, and Kelley’s teams can often swipe a ton of hidden yards from those particular people at the high school level.
That said, I’d love to see Kansas stop punting once it passes its own 40. It might not work, but playing conventionally definitely isn’t working. Why not give it a try?
For his part, Kelley is happy no one else seems to want to adopt his strategy. And because he’s incredibly efficient, he answered Matt’s question much more succinctly than I did.
I am glad they don't Matt😊. We keep having an advantage if they don't.— Kevin Kelley (@coachkelley1) April 10, 2018
From Doug: Can anyone realistically win the Big 12 besides Oklahoma? Why?
The Sooners have won the past three Big 12 titles. That streak coincides with two circumstances—Lincoln Riley running the offense and Baker Mayfield playing quarterback. Only one of those two things is going to continue in Norman, which means there probably will be an opening for another team or two to at least compete for the conference championship.
If Oklahoma’s defense doesn’t get better, then whoever wins the quarterback job (Kyler Murray or Austin Kendall) likely will have to play as well as Mayfield did to keep the Sooners at the same level. That’s a huge ask on the offensive end—a Heisman Trophy winner won’t be replaced immediately by another Heisman-caliber player. If Oklahoma’s defensive line is truly better, the defense might make a bigger contribution. If it doesn’t, it could open the door for another team.
TCU played the Sooners in the Big 12 title game, but neither meeting last year was close. The Horned Frogs are a bit of a mystery at the moment. They’ll have a new quarterback, a revamped offensive line and new starters in several key spots on defense.
I’ll believe in Texas when I see the Longhorns blocking well and winning consistently. Neither has happened in a long time in Austin, so we’re going to need to see proof of concept before any discussion of conference titles. Oklahoma State, meanwhile, is breaking in new starters all over the field. The Cowboys have won 10 games in four of the past five seasons, so Mike Gundy’s team deserves the benefit of the doubt. But it’s tough to explain why Oklahoma State might compete for the title without first seeing how these particular pieces fit together.
West Virginia is intriguing, though. Will Grier is one of the nation’s best quarterbacks. He’ll be throwing to David Sills V, who tied for the national lead in touchdown catches (18) in his first full season playing receiver. He’ll also be throwing to Gary Jennings, who caught 97 passes for 1,096 yards, and Marcus Simms, who averaged 18.9 yards on 35 catches last season. Meanwhile, future NFL tackle Yodny Cajuste leads a line that returns four starters. The defense took a big step back last season because of inexperience, but if Tony Gibson’s group can get a little better, the Mountaineers might be in for their best season since joining the Big 12. They also get Kansas State, TCU and Oklahoma in Morgantown this season.
Want another possibility? How about the team that beat both of last year’s Big 12 title game participants? Matt Campbell is entering year three at Iowa State, and the program is deeper than it has been in years. Quarterback Kyle Kempt, who helped beat the Sooners in Norman last year, was granted a sixth year of eligibility by the NCAA. He’ll be handing off to David Montgomery, a workhorse back who gained 1,146 yards last year. Kempt won’t have receiver Allen Lazard anymore, but he will have Hakeem Butler, who averaged 17 yards a catch last season. We’ll know early if the Cyclones intend to compete for the league title. By the end of the day on Oct. 13, they’ll have faced Oklahoma, TCU, Oklahoma State and West Virginia.
Oklahoma absolutely should be the favorite in the Big 12, Doug. But let’s not hand the Sooners the title just yet.