Chris Brown of the popular and insightful "Smart Football" has talked with us several times on his site and via twitter, so we asked and he agreed to work with us to better understand the Air Raid offense employed by Dana Holgorsen and the West Virginia Mountaineers. Since he's covered the system profusely, and we'd just be linking his articles anyway, we thought it better to just work together. Chris fielded our questions and provided excellent information on the offense along with WVU's head coach and its personnel. Please head over to Smart Football to view information on the Air Raid offense and all things football, as this is a very, very good place that provides excellent information in a format that is easy to understand. Again, we would like to thank Chris for taking the time over the Holiday Season to answer our questions with such quality, detailed responses.
Here are a few introductory notes from Chris before we get into the Q&A portion:
Like Chad Morris at Clemson, Dana Holgorsen immediately improved West Virginia's offense Also like The Chad, that improvement was not always universal or consistent; West Virginia's offense is far from a well-oiled machine, and has been prone to fits of turnovers, three and outs, and mental mistakes. But, to me, the improvement is real. I have lengthily recounted Holgorsen's background and the evolution of his attack for Grantland here, but, in brief: Holgorsen, who played Hal Mumme nd Mike Leach at Iowa Wesleyan, is a disciple of that "Airraid" offense, which Mumme and Leach put together based on the old Lavell Edwards BYU passing game. He coached under Leach and Mumme at Valdosta St. then installed the offense at Mississippi College and Wingate before joining Leach's staff at Texas Tech. Like Dabo Swinney he was a receivers coach, coaching, among others, Wes Welker. After eight years under Leach (including three with the title "co-offensive coordinator" or "offensive coordinator," when he was Leach's main eye-in-the-sky and play recommender, though of course Leach ultimately called his own plays), he wanted to install his own brand of the Airraid and got his opportunity under then Houston head coach and now Texas A&M HC Kevin Sumlin. In 2008 and 2009 his offenses at Houston, led by quarterback Case Keenum, exploded, finishing second in total offense in 2008 and first in 2009. The 2009 Houston squad boasted victories over three BCS opponents: Oklahoma State, Texas Tech (where Holgorsen beat his former mentor on a Case Keenum scramble in the waning minutes), and Mississippi State.
That season-particularly the victory over Oklahoma State-wound up being an audition for Holgorsen, as Mike Gundy hired him to install his offense at for the Cowboys after a disappointing 2009. The 2009 Cowboy offense fell from averaging over 36 points a game within the conference in 2008 to around 25 in 2009, saw the graduation of its quarterback, and suffered from the departure of number of other skill players including Dez Bryant. 2010 turned out to be a great year for Oklahoma State (surpassed only by 2011), as the Cowboys finished third in total offense and third in scoring under Holgorsen and won 11 games, the most Mike Gundy had ever won in his time as head coach. This was good enough to get Holgorsen hired as "coach-in-waiting" at West Virginia, where he was to turn around the offense as offensive coordinator for a year under Bill Stewart. I won't recount this history but suffice to say it didn't work out and Holgorsen became the head coach after spring practice but before fall camp, and has been both installing his offense this year and learning the ropes as a first-time head coach at the same time.
During his time at Houston, Oklahoma State, and now West Virginia, it's important to realize that Dana Holgorsen is not running Mike Leach's offense. Many of the principles including practice philosophy and a pass first mentality are very similar to Leach's approach. However, Holgorsen is more than happy to run the ball if an opponent is giving them that opportunity and, as a result, his teams don't quite look like Leach's. As I'll touch on below, he doesn't use the super-wide splits because his run game is based around the inside and outside zones and he wants double teams at the point of attack, and, rather than run 50 5-step dropback passes a game, Holgorsen wants to use those same concepts but mix them up with play-action, movement passes, and other wrinkles. Issues on the offensive line (that will be discussed later on) forced him to get away from the Airraid dropback game staple concepts, more out of necessity than desire. As a preface, the clips of every Geno Smith pass against LSU gives a good flavor of the good and bad for this team: You can see Smith at his most accurate and also at some of his most erratic (he's generally been good about not throwing passes that could be picked); you'll see how dangerous Tavon Austin is with the ball in his hands and how consistent Stedman Bailey was at beating even top defensive backs like Mo Claiborne and Eric Reid, and you can also see how this team can shoot itself in the foot and how the offensive line can get overwhelmed at times and give up hits and pressure.
STS Question: Dana Holgorsen began using the diamond (or inverted wishbone) formation last season at Oklahoma State and is using this formation along with variations of this formation at West Virginia. Do you think this is a permanent installation or just an intermediate until he can revert back to a pure Air Raid attack? Did this evolve strictly as a power run formation (i.e., goalline and short yardage) or to provide some variation within the pass happy Air Raid attack?
The evolution of the three-back set is interesting and I think it's a good illustration of why it's a mistake to look at what Dana wants to do as simply the "Mike Leach" or Hal Mumme/Leach "Airraid," at least in the pure form as they did at Texas Tech. First, while Dana definitely wants to throw the ball, he's not nearly as much of a purist as Leach or Mumme are (and remain). In Holgorsen's last year at Texas Tech, in 2007, Tech was second in the nation in total offense and yet averaged an amazingly low 59 yards rushing per game. The next season at Houston, Holgorsen's offense averaged over 160 yards rushing per game. Last season Oklahoma State averaged around 175 yards a game on over 5 yards a carry. Again, Dana isn't trying to be a flex option rushing team but, unlike Leach, if you play with two deep safeties and drop 8 guys into coverage -- or bring funky overload blitzes to one side or another -- Holgorsen wants to be able to run the ball to take advantage of that. If you show a single deep safety or eight-man fronts, he's more than happy to throw the ball all day long.
The three back grew out of this philosophy and was a response to the personnel they had at Oklahoma State. They had an all-conference fullback and a couple of others on the roster, and they had Kendall Hunter, who is now the change-of-pace back for the 49ers, as the featured back. They also liked the kids they had on the outside, one of which just so happened to turn out to be Justin Blackmon. In typical Airraid fashion, the theory of the three-back stuff is to actually throw the ball -- to get one-on-one matchups to the outside if the defense loads the box and puts a single-safety deep. Holgorsen often simply lets his quarterback signal whatever route he wants to those outside receivers.
The other reason it's a concept that will continue to be seen in Dana's offense is he really likes fullbacks, much more so than tight-ends (which I disagree with, I prefer TEs). His run game is very simple (more on that in a moment as well) but he feels like he can use fullbacks in more varied ways than tight-ends, and, more practically, can recruit them easier. So if nothing else Holgorsen's three-back set with a deep pistol back is really just his spin on someone else's basic I-formation with a tight-end, fullback, and deep back.
West Virginia has gotten away from this set for a couple of reasons, one being depth at fullback (they came into the year with two decent ones but have had some injuries there all year), inconsistency at running back, and, most importantly, poor line play.
The thought process for the set is as I described it above: if the defense respects run, you can throw one-on-one routes to the outside all day. However, if they play your heavy set with two deep safeties, you better be able to have your seven block their seven, and that's simply not something West Virginia has been able to consistently do. Syracuse played this set with a quarters coverage, doubling the outside receivers, and West Virginia hasn't been able to consistently block people or have their runner make guys miss, which were both things Holgorsen benefited from at Oklahoma State.
STS Question: West Virginia has shown patience at times, hitting quite a few screen nd swing passes in addition to the intermediate passes and the home run ball. What would you say is a typical pass distribution (short, medium, long) for an Air Raid offense and do you believe that Holgorsen historically sticks with these trends or bucks them?
I think Holgorsen's pass distribution is different at the margins from the typical Airraid team. They all, including West Virginia, throw a ton of screens and throw a lot of short-to-intermediate passes, primarily off of their dropback game. What distinguishes Dana's attack is this (start around the 50 sec. mark):
That stuff is seven-man gap or slide protection, often play-action, and downfield stuff. Some of the other Airraid teams have evolved to more play-action, but maybe the best way to describe what Dana is doing is he's come first circle: The Airraid concepts came from the pros via BYU (BYU assistant Doug Scovil got them from Sid Gillman, and Hal Mumme got them from BYU), and now Dana is putting his own spin on the pro-style downfield play-action game, often with one-on-one routes to the outside. It's only 25% of the offense or so, but that's a big change from what you'd see from Leach or the other true disciples.
STS Question: We expect to see a lot of "triangle concepts including some combo snag routes. If you were Clemson defensive coordinator Kevin Steele, how would you try to stop this multi-level attack and where would this leave the defense vulnerable?
The strength of West Virginia's offense is Geno Smith throwing to Tavon Austin and Stedman Bailey, but it's the offensive line and running backs that have dictated Dana's playcalling. Against the elite defensive lines in the Big East (don't laugh, in something of a fluke the Big East actually has a solid number of NFL draft picks across the defensive lines this season), WVU's line really struggled to block on both run and pass plays. The Mountaineers struggled against both fronts that blitzed a lot and those that didn't (they generally haven't had too many mental breakdowns with free blitzers, but instead more of the toughness or physical mismatch varieties). Further, WVU's leading rusher -- and often pass protector -- is a 160 pound true freshman two-star running back recruit named Dustin Garrison: not optimal for helping with blocking either. The kid has excellent vision, which is why he won the job over some veterans and higher touted recruits, but lacks speed and strength so tends to be most effective when the pass game is going. They have a bigger change-of-pace back named Shawne Alston who was hurt early in the year and managed to pick up 10 touchdowns, but he lacks burst which prevents him from taking over as the full-time guy.
Overall, if I'm Kevin Steele I start with personnel, not schemes: Who can hurt me and who can I exploit? It's clear that the outside guys like Bailey and Austin can kill me -- and they use Austin in a bunch of different ways, as I'll touch on shortly -- but the whole house of cards can fall in on itself if the line doesn't play well. My personal approach to WVU would be to stick to four and maybe the occasional unexpected five-man pressure (corner blitz, etc), keep a lid on the coverage with a couple of safeties deep and try to roll up my corners on Bailey and pay close attention to Austin. I wouldn't want to do too much in the way of blitzing unless I couldn't get pressure with my front guys. I also think scheme is probably less important than trying to show Geno Smith a lot of looks. Holgorsen puts a lot of pressure on him to signal specific routes and plays, and while he's been solid it's clear that, in his first season in the offense, Smith doesn't have the mastery of the offense of, say, Graham Harrell or Case Keenum or someone like that.
STS Question: I've noticed most Air Raid attacks use rather wide splits up front, presumably to address pressure from the end. This appears to leave the interior susceptible to pressure from a blitzing ILB and/or really good play from a DT. How do Air Raid teams mitigate this inside threat created through formation?
While it's true that Leach used the wide splits -- and I think WVU's trend slightly larger than average -- Holgorsen doesn't use the really wide splits. The purpose of the wide splits was to, as you say, force the rush to come from the outside. It also creates throwing lanes for all the short crossing routes Leach loved. "Vertical sets" also aided Leach's attacks. This is a technique where, essentially, the offensive line backpedals several yards before tightening up to stop the pass rush. By doing this the line can see all the stunts happen in front of them and help each other out more easily. It does seem to create a bunch of inside blitzing lanes but Leach's teams never seemed to have issues with these due, in large part, to these vertical sets. (One unappreciated aspect of them was how useful they were for receiver screens -- linemen were practically already lined up wide enough to start blocking for all those receiver screens Leach loved.)
The problem with the super wide splits is how useless they are in the run game for anything other than inside draw plays and certain outside runs if the defensive line tried to pinch down inside. Because Holgorsen's run game is based around the inside/outside zone (Hal Mumme famously for a long time had a total of two run plays: the draw and the lead draw) and he wants to create double teams with his linemen, he doesn't use the wide splits with the vertical sets but instead uses more traditional splits with a more traditional "kick-slide" technique.
I will say that some of this plays into the poor line play by WVU's line. Some of it was it simply was not a focus of recruiting after Rich Rodriguez and Rick Trickett left Morgantown, but it's also a matter of style. Rodriguez and Trickett wanted small, slim offensive linemen who weighed around 275-280 pounds and could rip and run because the base run play for Rodriguez was the outside zone. The last staff at West Virginia didn't have much of an identity (and the line didn't do well for them). Holgorsen, on the other hand, wants bigger guys like Leach had. Maybe not quite as big (after seven or eight years of recruiting Leach, with his big splits and vertical sets, often fielded lines full of guys 330-350 pounds), but you want guys more in a traditional NFL mold who can throw up a wall on pass protection to hold their ground and get more of a vertical push on the inside zone, as opposed to the Rodriguez/Trickett focus on outside zone leverage and reach blocks.
Finally, I think all the stuff I talked about above is great -- marrying the Leach Airraid with a hybrid pro-style zone run game with seven-man protection and deep play-action passes to get big plays -- but it all puts a ton of pressure on the offensive line to stop blitzers, open up run lanes, and give Smith time to find guys downfield.
STS Question: Are there any major items pertaining to the Air Raid offense that Clemson fans should focus upon while watching the Orange Bowl?
Regarding the offense, I'd focus on the things Holgorsen does to mix and match the base Airraid stuff with different looks and wrinkles. He does a great job taking the staple concept and adding motion, a few formations, play-actions, screens, combination or packaged plays. I think one thing both Chad Morris and Holgorsen do -- and that I expect you'll see even more from Morris as his offense evolves and as Clemson becomes more comfortable with them -- are the packaged run/pass or pass/screen type plays that Holgorsen has really been at the forefront of.
More specifically for this game, I'd begin the game focusing on the trenches. Like most football games, if West Virginia can't keep Clemson out of its backfield on run or pass plays, Bill Walsh himself couldn't scheme his way out of it. And if that is the case I think it will be really difficult for West Virginia to score enough to keep up with Clemson's offense (which I expect to have success against West Virginia's defense).
STS Question: I've noticed that Holgo doesn't run the classic Mesh play which I've considered a staple Air raid concept, why do you think that is?
Mesh, while a great play, is extremely expensive. The reason it's so good for the Airraid guys is they treat it like option teams treat the veer: You don't "dabble" in it, it defines what you do and you work on it every day. It's also a very specific concept and is susceptible to pattern reading -- with all those plays running to the flat or on crossing routes, the defense knows it's coming -- which is why to use it well you not only have to run the play but all the tags off of it (adding wheel routes, or faking the mesh crossing routes and running "whip" routes back to the sideline, etc). Because Dana does this other stuff in the run game and with the downfield passing game, he just doesn't use it enough to make practicing it worth it. They used to use it some at Houston but I think they just got away from it and I haven't seen West Virginia run it yet.
STS Question: Why so many run/pass options when Smith isn't an instinctive runner?
Well, I may have been unclear when I said run/pass options. I don't mean "run/pass option" in the announcer sense where they always say they want a "run-pass option" on the goal line or two-point play, like a bootleg or rollout play. Instead what I mean is that on running plays Holgorsen almost always, and essentially always in the red zone, let's his quarterback throw a quick pass to an isolated receiver. This is often called the "look" or "smoke" concept, and is extremely common in the NFL. They typically leave the run play on and the quarterback just steps away from the rush in the gun and throws a quick one to the receiver. Oklahoma State did a fantastic job at this concept this season; they actually built on it and made it an even bigger focus under Todd Monken after Holgorsen left, likely because Brandon Weeden liked it so much and was so good at it (and with Justin Blackmon on the outside it really presented issues for defenses). This has been OK for West Virginia this year, but the problem is that Geno Smith is not a very good fade passer, as evidenced by his poor completion rate in the red zone. It's possible Holgorsen was spoiled because Weeden and Blackmon were preternaturally good at it, but it's Smith's call whether to throw the fade, slant, or hitch on a called run play and he hasn't done a great job with his accuracy on those.
The concept, however, is a good one, and as the year went on they focused less on throwing fades on these kinds of plays than throwing screens to guys like Tavon Austin at the QB's discretion, and this was more effective for them.
STS Question: Other teams have taken away the Bailey kid, and he and Austin are the prime weapons. How has Holgo adjusted in the passing game?
One of the ways Holgorsen is true to the Airraid is they lock in where guys line up, though again Holgorsen puts his own spin on it. Whereas Hal Mumme and Mike Leach always put "X" (the split end) to the left and "Y" (the tight-end or inside slot to the right), Holgorsen moves the two inside receivers ("H", i.e. Tavon Austin) and "Y" to both sides but locks in "X" (Stedman Bailey) on the left and his "Z" to the right (primarily Ivan McCartney but a few other guys for WVU this year; this is the position Blackmon played). The upshot is that on a large percentage of snaps Bailey and Austin both line up to the left, and as the season went on the production from the other two positions, Y and Z, really trailed off. This wasn't a problem for Leach in the later years because he typically had solid receivers who had played in the system spread across his team, but in his first year Holgorsen hasn't had that luxury and that has allowed defenses to basically take away the offense's left side. Now Bailey hasn't really been shut down, but so much of the offense runs through him and he's very much a straight line receiver so the only answer is for other guys to step up. Austin, despite being extremely explosive, is still a developing receiver and a large portion of his production comes not in the old Airraid style of just off dropback pass plays. They use him on a variety of shovel passes, screens, their shallow cross screen (receiver runs a shallow while a runningback blocks the flat defender to the side the shallow guy is coming from -- another cool Holgorsen play that I see NFL teams using these days), and so on.
This means that they should be able to get Austin the ball no matter the defensive scheme, but it also means he's not necessarily going to draw downfield coverage away from Bailey as Holgorsen would like. The upshot here is that other guys need to step up for them if they want to have a chance to put enough points on Clemson.
To sum up, I'll focus on a couple of statistics showing both the progress Holgorsen has made with the offense along with some of its major problem areas.
- Both overall and when limited to conference play, West Virginia's offense improved by a full yard per offensive play this season as compared to last, and approached that number for the best of the Rich Rodriguez years. (In 2011, WVU averaged 6.34 yards per play overall and 6.11 per play in the conference; in 2010 those numbers were 5.3 and 5.2, respectively.) To me, this is an important stat, and a yard is a massive difference -- there's no question Chad Morris improved Clemson's offense, and the Tigers' O improved by about .7 this year as compared to last.
- Under Holgorsen, WVU averaged 10 more points a game overall (35) and roughly seven more per game against conference only opponents, as compared to last season. This is a significant improvement and a testament to what Dana has been able to do, given that his offense fields essentially the same players as the 2010 WVU team minus arguably its two *best* playmakers, Noel Devine and Jock Sanders.
- One could also look at WVU's weak third down conversion percentage here (Holgorsen's offenses have always been good, but I focus on Geno Smith's abysmal 46.5% completion percentage in the red zone. Overall, Smith is 64.7% of his passes, and from his opponent's 39 to the 21, he's lethal: 71%, nearly 10 yards per pass attempt, and 8 touchdowns to no interceptions. But once in the red zone, things get haywire. One might suspect that this is a spread offense thing, but the numbers don't back that up: In 2010 at OKST, Brandon Weeden completed 55.5% of his passes in the red zone and his pass efficiency rating was almost 30 points higher than Smith's; in 2009, under Holgorsen Case Keenum completed 63.6% of his passes in the red zone (and had a passer rating almost *sixty* points higher than Smith's), and in 2007, Holgorsen's last season at Texas Tech, Graham Harrell completed almost 63% of his passes and his pass efficiency rating was roughly 55 points higher than Smith's. (All stats courtesy of cfbstats.com.) As previously discussed, some of this was probably an overreliance by Holgorsen on fade passes near the goal line (having been spoiled by the efficiency of Brandon Weeden-to-Justin Blackmon fade combination), but it's something Smith will have to do well in the bowl game if they want to win and will be important next season.
- West Virginia averaged 117.75 rushing yards per game this season, not a great number but, not abysmally low considering how high the passing stats were. But that doesn't tell the whole story, with an inconsistent offensive line, a new system, and a bunch of freshman and inexperienced runningbacks, the running game was just flat *inconsistent*. At Oklahoma State, Holgorsen had the luxury of handing off to a couple of really good backs: Kendall Hunter, who is now the change-of-pace guy for the 49ers, and Joseph Randle, a true dual threat who this season as the featured back had nearly 1,200 yards on the ground and 23 touchdowns on over 6 yards per carry. Consider this to get a flavor of how inconsistent WVU's run game was:
- In their top six games in terms of rushing output, the Mountaineers averaged 172.5 yards per game. They went 5-1 in these games.
- In their worst six games in terms of rushing output, the Mountaineers averaged 63 yards per game. They went 4-2 in these games.
- In their top six games in terms of rushing output, the Mountaineers averaged 172.5 yards per game. They went 5-1 in these games.
Those teams used different approaches: some were blitzing teams, others only rushed four, and so on. As I touched on above, if WVU's line plays well, WVU has an excellent shot to put up a lot of points. If they don't, and Clemson and Kevin Steele can keep a lid on the big plays the Mountaineers have been so good at this season, I think WVU will really struggle to be consistent enough on offense to keep up with Clemson's O.