Inside the Numbers: The Success of Senior Starting Quarterbacks

After playing a strange charade with the media all spring and summer, claiming that there was a quarterback competition between Devin Gardner and Shane Morris, Brady Hoke surprised absolutely no one when he announced that Gardner will be Michigan's starter against Appalachian State. Why would Gardner not be the starting quarterback in 2014? Sure, there are some who point to his 10-7 record as a starter, albeit foolishly, as evidence that he is not a winner or a Michigan-caliber quarterback. They say this despite that Gardner cannot play all 22 positions on offense and defense. They say this despite that he has led six career game-tying or lead-changing drives in the final five minutes of regulation. They say this despite that he just recorded one of the most productive seasons ever for a Michigan quarterback behind arguably the worst offensive line in school history. And they say this despite that Gardner will be a fifth-year senior in 2014.

Senior starting quarterbacks are a desired commodity in college football. It is believed that offenses generally are at their best when an experienced senior quarterback rather than an underclassman drives them down the field. This makes total sense on the surface. Seniors have been in the program for at least three years, understand the intricacies of the playbook, and are acclimated to the ins and outs of college football. Meanwhile, most underclassmen struggle to learn instantly all of the complex concepts being thrown at them. There are talented exceptions, of course. Look no further than the last two recipients of the Heisman Trophy, Johnny Manziel and Jameis Winston, both of whom won the prestigious award as redshirt freshmen. But they seem to be the anomaly, not the rule. So, with Gardner beginning his senior season this Saturday, this edition of "Inside the Numbers" researched whether the statistics support these assumptions about quarterbacks and their age and projected what this means for Gardner.

The Success of Senior Starting Quarterbacks

Other Michigan writers or statisticians have have wondered about a senior quarterback's value to an offense as well. Brett Thiessen, a contributor at MGoBlog, penned a post tackling this exact topic last summer. He collected offensive data for each FBS school from the 2011 and 2012 seasons in addition to information about the age and experience of the quarterbacks starting for these teams to create the following chart, which supports the aforementioned belief that offenses typically are most productive when led by an experienced senior quarterback:

[Brett Thiessen-MGoBlog]

Analyzing this chart, Thiessen concluded that teams commonly do not follow a winning formula when they start a true freshman at quarterback. On average, among all classes of eligibility, true freshmen provide the least amount of value to an offense and usually undermine the efforts of the entire unit. But quarterbacks tend to improve as they age and accumulate more starting evidence, as Thiessen's Points Above Normal (PAN) metric indicates. By the time they are seniors with some starting experience under their belt, they routinely are the most valuable quarterbacks to any offense. On the other hand, coaches should be wary of inserting senior quarterbacks that have zero career starts into the lineup. In most cases, they were not the starter previously mostly because they were not very good, not because they were just stuck behind a better candidate..

You may think that Thiessen's conclusions were mostly common sense. Maybe. None were particularly groundbreaking. But they still were significant. Thanks to Thiessen, there now is statistical evidence that offenses commonly thrive with an experienced senior quarterback taking the snaps, even if there always will be exceptions.

Michigan's Senior Starting Quarterbacks

Since Thiessen had mined the national data to measure the value of an experienced senior quarterback to an offense, this week's "Inside the Numbers" narrowed its focus to Michigan's senior quarterbacks only. National statistics may suggest that offenses customarily are at their best with a senior incumbent at quarterback, but this does not necessarily mean that Michigan has enjoyed the benefits of this phenomenon. So "Inside the Numbers" gathered data from the past 25 years of Michigan's offense to establish how often it had a senior starter at quarterback, how well those seniors performed, and how well the offense fared under their senior leadership.

Since 1988, Michigan has started eight seniors at quarterback:

To evaluate how well these seniors performed versus their younger selves and other younger quarterbacks at Michigan, "Inside the Numbers" dove into two individual metrics and one team statistic. Passer rating and individual total yards per play were chosen to examine individual production. Offensive points per possession was chosen to examine a team's production. What each represents will be explained below, but the numbers from each conveyed the same message: in the past quarter-century, Michigan's senior quarterbacks routinely have been among the nation's best despite demonstrating no development after their junior season, which in some cases was due to injury.

Passer Rating

Passer rating is a marvelous tool that measures a quarterback's ability to throw a football with precision and efficiency. What makes it so valuable is its formula, which incorporates four statistics critical to determining a quarterback's performance through the air: completion rate, touchdown rate, interception rate, and yards per pass attempt. Weighing each of these, the formula spits out a number which represents how productive and efficient of a passer a quarterback has been. One caveat is that passer rating does not account for a quarterback's ability to make plays his legs, but this will be addressed by the second individual metric: total yards per play.

Michigan's seniors ordinarily demonstrated that they were proficient passers, sometimes even quite exceptional. These eight seniors, all of whom had earned some starting experience as a freshman, sophomore, or junior, combined for a passer rating of 139.7 in their final season. These days, this would be only above average. But this combined rating was calculated using some data that is 25 years old. It would be unwise to compare it to ratings only from 2013, especially when advances in passing schemes have caused quarterback play and passer ratings to soar in recent years. Looking at a historical context, a passer rating of 139.7 is praiseworthy. It would have been the 24th-best nationally in 2007. It would have been the 15th-best in 1999. And, then, when one realizes that Michigan has had seniors lead the nation in passer rating (see: Elvis Grbac in 1992), it is safe to suggest that the Wolverines generally field one of the elite passers in college football when they start a senior.

Yet these Michigan quarterbacks typically failed to improve as passers during their senior season. They seemed to reach their full potential prior to their final year of eligibility. The combined passer rating for junior starting quarterbacks at Michigan was 142.0, which is a few points better than the seniors' 139.7. In that final offseason, these quarterbacks ordinarily did not progress as passers. If anything, there was a slight backslide more often than not.

Instead, Michigan quarterbacks made their largest leaps between their sophomore and junior seasons. On average, their passer rating spiked from a poor 128.0 as sophomores to a commendable 142.0 as juniors. This seems to be when everything just clicked for Michigan's quarterbacks. This was even the case for those that did not earn their first career start until they were juniors. Despite holding the clipboard for two years, these juniors fully capitalized on their first opportunity to be the full-time starter (see: Todd Collins and his 149.3 passer rating in 1993). But it was rare to see this type of growth from a Michigan quarterback between his junior and senior seasons.

Individual Total Yards Per Play

Individual total yard per play is the second statistic that will be used to supplement the findings in the foregoing section with regards to passer rating. Whereas passer rating measures only a quarterback's ability to throw the football, individual total yards per play takes into consideration a quarterback's legs. In some cases, the strength of his legs is being a run-first quarterback and tallying over 1,000 rushing yards in a season (Miss you, Denard!). In other cases, it is maneuvering around debris in the pocket to avoid sacks. Either way, how well a quarterback uses his legs is crucial to his performance, so it is analyzed here.

A trend seems to be emerging. As Michigan's starting quarterbacks developed in the early stages of their career, their individual total yards per play surged each season. Freshmen starting quarterbacks averaged a measly 5.60 total yards per play. This would not have even the cracked the top 100 nationally in any season from 2007 to 2013. But this figure jumped to 6.28 total yards per play for sophomores before springing to 6.80 total yards per play for junior starting quarterbacks at Michigan. This was a substantial increase. A different of 1.20 total yards per play may equate to a 600-yard difference for a quarterback over the course of the season. Those 600 yards may have turned certain losses a Michigan quarterback suffered as a freshman into wins as junior. And this was because he learned to be either more dangerous as a runner, comfortable in the pocket, or efficient as a passer.

But, once again, the improvement in a Michigan starting quarterback's total yards per play came to a screeching halt once he was a senior. Sophomores upped their total yards per play by 0.68. Juniors upped theirs by 0.52. But seniors generally could not replicate this growth. Instead, they posted 6.79 total yards per play, which was 0.01 fewer than what the juniors registered. This is just further proof that Michigan's senior quarterbacks either had no room left for development entering their final season or suffered an injury that hindered their development. So, in most cases, what Michigan fans saw from junior starting quarterbacks is what they saw from them as seniors.

Offensive Points Per Possession

The first two metrics explored the individual production of Michigan's starting quarterbacks, but offensive points per possession determines how well the offense fared with either a freshman, sophomore, junior, or senior as the starter. This is better than points per game because points per possession can separate the contributions of two quarterbacks from the same team playing in the same game and disregards defensive and special-teams touchdowns. This also is better than team total yards per play because points per possession accounts for points, which are somewhat important in the game of football, punts, turnovers, and failed fourth-down conversions. It must be noted, though, that there are some discrepancies in the numbers because no Michigan play-by-play information was available for any games prior to 1998. Although this means that chart below is not 100-percent accurate, the discrepancies are not significant enough to materially affect what the numbers are indicating.

Okay. It is official. There is a legitimate trend apparent in all three of these charts. Once again, Michigan's offense exhibited constant improvement in its ability to score as its quarterback aged early in his career. The offense mustered only a putrid 1.985 offensive points per possession when led by a freshman quarterback, but this figure sprouted all the way to 2.464 offensive points per possession when the offense was led by a junior. This difference is enough to add another 70 or so points to the scoreboard for the Wolverines throughout an entire season, which is significant. This should not be much of a surprise, though. As Michigan quarterbacks aged from underclassmen into juniors, they learned how to limit and correct their mistakes, which increased the Wolverines' odds of scoring on every possession.

But, oh look, there was little to no difference in the Michigan offense's ability to score when it was being quarterbacked by either a senior or junior. The progression stagnates. With a senior under center or in shotgun, Michigan averaged 2.523 offensive points per possession. Unlike the previous two statistics, this actually was an improvement from how the offense fared with a junior quarterback. But it was a minor increase. It is not an increase that persuades one to conclude that the Michigan offense should be expected to breakout with a senior quarterback at the helm..

What to Expect from Devin Gardner in 2014

All three of these offensive metrics--passer rating, individual total yards per play, and offensive points per possession--indicate that, although Michigan's senior quarterbacks were considered among the best in the nation, they performed quite similarly to how they did as juniors. There was no drastic change or improvement. There was no last-minute breakout season. Michigan knew what to expect from their quarterbacks by the time they had become upperclassmen.

This is why fans should not expect anything from Gardner in 2014 vastly different from what they saw him do in 2013. The good news is that Gardner was one of the best quarterbacks in the Big Ten last season. He demonstrated that he could be productive and efficient. He finished second and eighth in the Big Ten in total yards per game (286.9) and total yards per play (6.75), respectively. However, among returning Big Ten quarterbacks that will be healthy to play this fall, these numbers were the first- and second-best, respectively. He posted a passer rating of 146.07, which was second in the conference only to Ohio State's Braxton Miller. He also knew how to put points on the scoreboard, accounting for 32 total touchdowns--tied for the second-most in school history. And, even more impressively, Gardner achieved all of this behind an offensive line that allowed the most tackles-for-loss in the country. Simply, fans know that Gardner is a playmaker with both his arm and legs. This should not change in 2014.

However, Gardner is by no means a perfect quarterback. He displayed his fair share of flaws last season. And, if this statistical analysis has enlightened us about anything, it is that Gardner should not be expected to rectify most, if any, of these flaws as a senior. Gardner still will likely not go through his progressions while reading a defense, instead locking onto his first target and telegraphing his throw. He still will likely take too many risks in trying to make a big play when nothing is there, leading to a high turnover rate. He still will likely throw off his back foot when he eats pressure in the pocket, causing him to be inaccurate and throw passes ripe for an interception. This is not to say that Gardner has not practiced hard to fix these bad habits over the spring and summer. But, with Gardner entering his final season of eligibility, this is who he is as a quarterback.

Of course, Gardner could be an exception to the rule. There are always exceptions. Just because most of the previous eight senior quarterbacks at Michigan did not progress during their final season does not mean Gardner will share the same fate. Gardner is not Michael Taylor, Grbac, Collins, Brian Griese, Tom Brady, John Navarre, Chad Henne, or Robinson. He is his own player, capable of doing what those before him have not been able to do. Under new Michigan offensive coordinator and quarterbacks coach Doug Nussmeier, maybe Gardner eradicates all of his bad habits and turns in a Heisman Trophy-contending campaign. Maybe Gardner is able to improve as a quarterback if a young, inexperienced offensive line in front of him gels and provides him with more time than he had as a junior. Anything is possible.

But, if recent history has taught us anything, it is that Gardner, as a senior starting quarterback at Michigan, should be one of the, if not the, best quarterbacks in the Big Ten in 2014, even though he should continue to make the same mistakes that frustrated Michigan fans last season.