Sunday, May 29, 2011

Iowa Football Run/Pass Splits Since 2004

Now that I'm back from my trip to the former Yugoslavia, and my stomach is fully rested after consuming many liters of homemade sljivovica. It's time for me to get back into the swing of things. That means it's time for me to dig deep into some Iowa Hawkeyes numbers.  Probably way deeper than is normal for any sane person. But, hey, I had two weeks pretty much cut off from the world. I'm well rested.

Anyway, for my first post back, I went back to 2004 to look at the Iowa Hawkeyes' run/pass tendencies under Kirk Ferentz. I only went back to 2004 because that is as far back as the Hawkeyesports.com's boxscores included the "play breakdown" section, which made this exercise easier than it would have been otherwise.

My main interest in looking at this, was that I wanted to see how Iowa under Ferentz, adjusted their offense based on what they have to work with from year to year. For instance, what does the Iowa play calling look like when they have a running back like Shonn Greene, as opposed to when Sam Brownlee is the man taking the majority of the carries?

A main limitation of this study, is the fact that not all box scores pointed out the difference between a quarterback who dropped back to pass and instead ran with the football, from a quarterback who took the snap and ran with the ball by design. Some box scores made this distinction, but most did not, which left me using things like down and distance to determine whether or not it was a called pass play or run play. For example, 3rd and 9, and Jake Christensen runs for 8 yards. It probably was not a designed run. Instead, it was probably the protection breaking down, and the quarterback scrambling for however many yards he could muster. The box score would count that as a running play, but I would call that a pass play, considering the coaching staff intended for that play to have the ball in the air. This makes my numbers look different from the numbers you would see if you looked up the run/pass splits from Hawkeyesports.com. My numbers aren't perfect, as I had to guess, but they are more accurate than the official websites numbers are.

Anyway, let's see the results.

2004


As you no doubt remember, 2004 was the season of the exploding knee's for Iowa running backs. Sam Brownlee finished the season as Iowa's leading rusher, as he put up a record setting 227 yards on 94 carries. As a result it, should come as no surprise to Hawkeye fans, that Iowa was more likely to pass that year, on every down except for second and less than 3 yards. Overall, Iowa's play selection was 58.17% pass plays on the year. Easily, the highest number over the last seven seasons under the Ferentz regime.

2005


The next year, Albert Young burst on the scene, putting up what would be his best year, as a redshirt sophomore. With a real rushing attack, the Ferentz led Hawkeyes went from calling passing plays about 58% of the time, to calling them on 53.33%. The Hawkeyes were more likely to call a running play on every down except for second and long (7 or more yards) and 3rd and medium (4-6 yards) or longer, and fourth down.

2006


The 2006 season saw Iowa go back to a more passing heavy attack. Iowa called a pass play on 56.51% of their plays that year, which was the third highest total from this study. With Albert Young and Damian Sims returning at running back, Ferentz probably wanted to rely more on them. However, with a Senior quarterback in Drew Tate, and games that were fairly close in score (Iowa's average margin of victory fell from 10 points the previous year, to 5.31 points in 2006), Iowa had to rely more on the pass in 2006.

2007


The 2007 season was a nightmare. To put it mildly, Jake Christensen had some accuracy issues (53.5% completion percentage). To put it not mildly, the only way you were likely to catch a pass from the southpaw quarterback, was if your last name was "Turf." Not to mention, that Iowa's offensive line was terrible that year, and you see why Iowa called a passing play on 56.74% of their play calls (second most since 2004, behind only the Sam Brownlee year). Ferentz had Senior running backs in Albert Young and Damian Sims to lean on, but the aforementioned offensive line problems, made the running game almost non-existent. With the games once again being much closer, even than they were in 2006 (average margin of victory of -0.58 points), Iowa once again leaned more heavily on the pass in 2007.

2008


2008 was a different story, however. We see that Kirk Ferentz realized he had a weapon. Enter, Shonn Greene. The 5'11" 235 lb bowling ball, gave Iowa the best weapon in the Big 10 that year. As a result, Iowa ran the ball on 54.47% of their plays that year. Only one of two years since 2004, in which Iowa called more running plays than passing plays. Basically, the only time Iowa wasn't more likely to run the ball was when they were on their third down and had more than 4 yards to go. Otherwise, they were giving the  ball to #23, and they didn't care who knew.

2009


2009 was a pretty average season in terms of play calling. Iowa ran the ball on 46.31% of their plays, and passed the ball on 53.69% of their plays, which is pretty much the average run to pass ratio for the Hawkeyes since 2004. Iowa had an underwhelming rushing attack in Adam Robinson (4.6 yards per carry) and Brandon Wegher (4.0 yards per carry), but they also had a Redshirt Junior Quarterback named Ricky Stanzi, who had shown an affinity for giving the ball to the other team. So, it doesn't come as any real surprise that Ferentz didn't really stray from his seven year average that season.

2010


2010 was interesting to me. Iowa was basically 50/50 balanced in 2010, calling a run play on 50.57% of their play calls. But, why? Iowa's running game was below average last year. They averaged 4.30 yards per carry, while the Big 10 average was 4.58 yards per carry. Adam Robinson had a below average 4.4 yards per carry, and Marcus Coker (5.5 yards per carry) didn't really establish himself until later in the season. Instead, Iowa had a fifth year quarterback in Stanzi, who had proven he had cut down on the turnovers, and showed an affinity for finding his receivers in the endzone. When you realize that Stanzi was throwing to two of the best wide receivers that had ever put on a Hawkeye uniform, it makes it a little curious as to why Iowa didn't throw the ball more?

Overall


Based on the the past 7 seasons, Iowa has been more likely to run the ball on first down, second down and short (1-3 yards), second down and medium (4-6 yards), and third down and short (1-3 yards). The Ferentz led Hawkeyes, have been more likely to pass the ball on second and long (7+ yards), third and medium (4-6 yards), third and long (7+ yards) and fourth down.

Overall, the Hawkeyes are pretty balanced, but have called a pass play on 53.38% of their play calls, while running the ball on 46.62% of play calls. As we've seen, Ferentz usually goes with the weapons he has on offense. In 2004, when Iowa couldn't keep a scholarship running back healthly, Ferentz called a pass play 58.17% of the time. In 2008, when Iowa had a workhorse running back in Shonn Greene, Ferentz ran the ball 54.47% of the time.

It's no surprise that Iowa's play calling would depend on the weapons the offense has at its disposal. However, I also wanted to look at how margin of victory affected the Hawkeyes' run to pass ratio. Below is a chart plotting every game since 2004, based on the Iowa's margin of victory and their run to pass ratio from that game.


There is clearly some relationship. According to R squared, the point differential of the game, can explain about 35% of the variance. The break even point (where the run to pass ratio equals 1) is predicted at the margin of victory of 10.24 points. An R squared of 0.35 is not very predictive, but it does still make sense. If a game is decided by less than 10 points, then Iowa probably had to pass the ball more to try to put the game out of reach. However, if the game is decided by more than 10 points, there is a good chance that it was basically out of reach with a good amount of time left, which means that Ferentz (who will notoriously run the clock out once he gets even the smallest of leads) probably was able to sit on the ball, and rely on his defense to not give up the lead.

Summary
 
So, what was my point in this really long post? Well, first off, Iowa is almost always going to be more likely to pass the ball more than run it over the course of a season. That is because third down is such a passing down that it skews the overall numbers toward the passing side. However, on first and second down under Kirk Ferentz, Iowa is more likely to run the ball.

Secondly, Kirk Ferentz does adjust to the players he has on the field. If he has no running game, he will throw the ball. If he has a star running back, he will feed him the ball, and it doesn't matter if the defense knows the run is coming. Basically, it doesn't matter if you know what is coming, because you still have to stop it.

Lastly, and unsurprisingly, how close a game is, affects Iowa's play calling in some way. The R squared, insisted that the point differential of a game, explained about 35% of the variance in Iowa's run to pass ratio. Not very predictive, but still a pretty large player in how often Iowa runs or throws the ball. In other words, it predicts that about 10 points is the break even point. Anything less than 10, and the game was too close for Ferentz to run the ball a majority of the time. Anything over 10 points, and Ferentz is running the ball, and also the clock down.

This upcoming season will be fun to pay attention to, as Iowa has a very good looking running back in Marcus Coker returning. They also have a new quarterback with two starts under his belt. If Coker and the offensive line are as good as advertised, this could be another season where Iowa runs the ball more than 50% of the time. However, depending on a multitude of things: How good of a quarterback is Vandenberg? How good is Iowa's defense at keeping opponents off the board? Can Iowa keep the angry running back hating God away from Coker?  We might be in for a lot of Vandenberg to McNutt.
Either way, I can't wait to find out. September 3 can't get here soon enough.


Follow Me on Twitter: GoHawks1123

3 comments:

  1. Great work. I was wondering if you have any numbers for yardage gained per play type/down. There's the idea from economics (and just common sense) that you should ideally get the same average yards per run and per pass. If you're getting more from running, on average, you should tilt the balance more toward the run until the defense adjusts. And vice versa for passing. Otherwise you're just leaving yards on the table, so to speak. Would be interesting to see how Iowa stacks up by that criteria.

    ReplyDelete
  2. jamerchant,

    Thanks. Unfortunately, I do not have those numbers, as gathering the numbers for this post took quite a while.

    That economics idea makes sense. The 2010 season is the only one where I was left scratching my head about the playcalling. I may have to gather those numbers later on when I have more time, at least for the 2010 season. Thanks for the idea!

    ReplyDelete
  3. That small R^2 value has my teeth gnashing. Run this with R (the stat program language) and determine/disregard outliers.

    ReplyDelete