The 2013 NFL Draft quarterback class is considered void of blue-chip talent by any and all draft experts. Headline names like Geno Smith and Matt Barkley have been picked apart and scrutinized from almost every angle possible regarding criteria such as work ethic, leadership skills, decision-making, and throwing accuracy. After listening to countless television and radio personalities make definitive statements on players they may or may not have actually studied in depth, I decided to take a closer look at most of the quarterbacks of the past several seasons who were either drafted, projected to be drafted, or that I took a particular interest in such as Penn State’s Matt McGloin.

Focusing solely on each quarterback’s third down throwing performance in college, a couple of things about these prospects perceived to be cloudy and uncertain becomes rather clear and very much tangible. I measured every player on their ability to successfully complete the objective on third down pass attempts, that being the play resulting in either a touchdown or a first down.  This metric, while not perfect by any means,  accomplishes the task of putting college football players in similar environments which can be difficult due to the disparities that exist in quality of opponent and style of play.  Third down pass attempt performance does a good job of getting the quarterback out of their comfort zone and allows for clear-cut parameters to be set up: did you succeed or fail in your attempt to gain a first down or touchdown?

The most important column in the graphic below is “WT_SR%” (third from right) which is the success rate on third down of all the quarterbacks listed in the leftmost column.  The “WT_+/-%” column (second from right) is the value above or below the average success rate of all players evaluated.  Data for this post is courtesy of cfbstats.com and sportsreference.com.

If the Buffalo Bills wanted a quarterback from this year’s class, they very well may have picked the right one in E.J. Manuel. The “WT_SR%” column I mentioned previously shows the success rate on third down for each player in their collegiate career.  The column to the immediate right is the percentage above or below average relative to the 38 quarterbacks observed for this piece.  The best value in the draft may be Landry Jones, a four-year starter at Oklahoma who showed significant improvement in his senior year.  As you can see in the subsequent data, this fits the Colin Kaepernick progression model with both players being four year starters who improved exponentially in their senior campaigns.  The complete table for most of the notable quarterbacks eligible for the draft from 2008 through next year’s projected class immediately follows this paragraph.  Feel free to play around with the 3 data tables found below.  The sheet #1 table shows all data compiled for this project.  Sheet#2 is a more complete version of sheet#3, which features each quarterback’s performance for each of their seasons as the primary starter for their respective college teams.  I’ve also included on sheet#2 the NFL winning percentage in games started and DVOA (percentage value above or below average) for all players with at least a handful of professional games started.  The default screen shown is sheet#3 and is sorted by the highest overall success rate measured, belonging to 2014 NFL Draft – eligible Johnny Manziel of Texas A&M.

Of the 21 quarterbacks on the list who have started at least a handful of games in the NFL, 12 performed similarly in the pros as they did in college.  Russell Wilson, who was significantly undervalued throughout the 2012 NFL Draft process, had the single best season observed of any quarterback in 2011 with Wisconsin.  Outside of Robert Griffin’s poor outlying performance on the above list, sorted best to worst total collegiate success rate, the bottom half of the table does a good job of identifying disastrous NFL prospects including Blaine Gabbert, Jimmy Clausen, and Ryan Lindley.  This does not bode well for NFL hopefuls Matt McGloin, Tyler Bray, and Tyler Wilson.  One wonders why college standouts like Sam Bradford and Colt McCoy have fared not so well in the professional ranks to this point.

Think of this entry as a starting point in attempting to measure the seemingly intangible abilities of all football coaches, with “intangible” being the very things which would set them apart from each other if all were dealt the same hand in terms of the talent level of their players.  In other words, which coaches get the best results on the scoreboard given what they have to work with season after season…and which coaches don’t.  All data is courtesy of sportsreference.com.

Methodology

The formula I created for measuring NFL head coaches is broken down into two main components:

• Team SRS Rating
• Team Net Yards Gained per Play (adjusted for strength of schedule)

The team SRS (simple ratings system) rating will tell us how well a team performs and by definition is a team’s average margin of victory over a presumably average opponent. It is preferable to team won-loss record for this exercise due to the misleading win totals, for better or worse, that can occur in brief 16 game seasons.  The 2004 Atlanta falcons, for instance, went 11-5 during a regular season in which they outscored their opponents by a total of 3 points and posted a below average SRS rating of -2.2.  The point differential and SRS rating are indicative of a 7 or 8 win team which given a league average schedule and one or two fewer made field goals would have been the case.  To clarify the meaning of the SRS rating take the best team measured in this study, the 2007 New England Patriots, who outscored the average team (SRS rating of 0) by 20.1 points per game and had an SRS rating of 20.1.  The worst team observed was the 2009 St. Louis Rams, who were outscored by Joe Average NFL team by 17.4 points per game for an SRS rating of -17.4.

A team’s net yards per play will be used to determine the talent level of each individual team and is calculated simply enough by subtracting their average yards allowed per play defensively from their yards gained per play offensively.  The best team over the past ten NFL seasons in this respect was the 2010 San Diego Chargers at +1.5 yards per play while the worst was the infamous 2008 Detroit Lions at -1.7 yards per play.  In order to strengthen the correlation between team SRS rating and Net Yards per Play,  I adjusted the “Net Yards” factor for each team’s strength of schedule.  For instance, observe the 2012 San Francisco 49ers adjusted net yards:

Year     Team       Coach              A           B           C                D           E              F

The 49er’s net yards per play is represented by A-B = C.  Column D is the strength of schedule component with a positive value representing an above average strength of schedule.  In this instance, I am going to increase the 49er’s Column C total  to reflect how they would have fared per play against an average opponent with a column D value of 0.  This new value representing adjusted yards per play is found in column F.  I used one tenth of the schedule strength component, represented by column E, due to it leading to the strongest correlation between SRS rating (not shown in the above example) and Net Yards/Play.  In other words, it most effectively extracted both of the main components out of our final objective of a truly “intangible” rating for all coaches.  This calculation, which represents the entire denominator in our final coach rating, is equivalent to… (A-B) + (D/10).  That’s all there is to it.

Now that the makeup of the denominator has been established, the complete evaluation of NFL coaches can be established as follows:

                                          SRS Rating / Net Yards Per Play

The final step is to figure the best way to represent the two components as to give the proper emphasis to each one.  I did this by scaling all the individual SRS values (320 in all) from .200 to .800 with .200 representing those pitiful 2009 St. Louis Rams mentioned above and .800 representing the 2007 Patriots.  The same idea is behind the scaling of the Net Yards per Play values from .300 to .700.  In this case the worst value of .302 belongs to the 2009 Cleveland Browns and best value of .700 belongs to the 2012 San Francisco 49ers.  I adjusted the final value so that an average coach would have a rating of precisely 1.  With the heavy lifting done we can review the results from the last ten NFL seasons.  The list is sorted by each coach’s percentage above or below the average rating, with the highest rated individual, the Falcon’s Mike Smith, listed first.  I’ve also supplied the number of seasons of data for each coach that were included in this piece.