Scout Notebook 9/8/10 – Completion % vs Accuracy

Welcome to the first edition of the Scout Notebook! At NFLMocks, we wanted to find a way to discuss in more depth how we analyze prospects. Often times people fall in love with the stat sheet (I blame fantasy football, but its too much fun to not play!) or use it as a crutch to support a player. Stats, however, clearly aren’t the only thing it takes to translate to the NFL. Just check the top ten at of any of the individual NCAA records and see how many NFL stars are there, especially the quarterback records. Quarterback is one of the toughest positions to scout and the process for evaluating them is full of conjecture and misinformed opinions because there is so much bias towards stats. Stats can be used in evaluation, but they still have to be applied correctly. Michael Rosenberg of SI.com recently wrote an article where he thrice tried to correlate stats with projected NFL success:

“Accuracy” Locker completed 58 percent of his passes last year — pretty good, but not special. Yardage? He threw for 2800 last year, 33rd in the FBS. Leadership? Locker has yet to post a winning record.”

All three statements use stats incorrectly to identify qualities in a quarterback. He pits accuracy against completion percentage, pure statistical data to future on field performance, and holds intangible qualities directly effect win/loss records. The worst offense would be linking leadership, an intangible quality, to wins on the field. Leadership can only take talent so far and talent can win without it on the field, but we can save that for another time. Today we are going to start at the beginning of this misguided statement and talk about how Accuracy and Completion Percentage are different.

Much like our initial example, these are simple examples of a bigger picture, but I hope the first thing you notice is that none of the three passes ever have a chance to hit the receiver in the numbers because the receiver  has his back to the quarterback as they are sprinting down field. So now our simpler example of using the receiver’s jersey numbers as our target has changed because the target has moved based on the play and the coverage. Because each of the receivers are able to get behind the coverage, throwing the ball behind them will only mean trouble and if the throw leads the receiver off of their route, it gives the defender a chance at the ball. In each case the best (a.k.a. the most accurate) place to put the ball is in front of the receiver and towards the sideline. If the throw is too far ahead of the receiver, at least no one has a chance at it. If it falls just right or slightly short, if it isn’t on the sideline side of the receiver, the defense again has a chance to make a play. In the name of equality, here are some accurate passes that hurt the quarterbacks completion percentage.

So we have seen how completion percentage can be affected by the players on the field, lets show how it can be affected by scheme. Colt McCoy was a favorite amongst fans last season. Often he was cited as an accurate passer with his high completion percentage as chief evidence of that fact. However, when you watch their games, the Longhorns were running a spread run and shoot system additionally using the threat of Colt scrambling to keep the defense near the line and open up shots down the field. It’s very similar to what June Jones runs everywhere he goes which is usually discredited come draft time for being a system that inflates quarterback talent. The only real difference is the talent levels between Texas and schools like SMU and Texas Tech. Here are the passing stats for McCoy in his senior year vs those of the SMU Mustangs (Jones’s current team) and Texas Tech Red Raiders (a heavy pass version of run and shoot) who ran similar systems.

The way the system works is by getting the ball to the receiver quickly on short slants, curls, comebacks, and crossing routes and let them get up field and keep the defense up towards the line of scrimmage. This is usually a function of the run game, but in these systems there is a bigger emphasis on passing so some of those passes must replace the effectiveness of a run game. This inevitably leads to a high “pitch count” or high number of pass attempts. McCoy had 33.6 att/gm, SMU 36.3, and Tech 51.5. McCoy’s running ability is most likely what keeps him from having as many attempts as these schools because many of his rushing attempts came on plays calling for a pass, but he decided to pull it down and run. McCoy by himself averaged 9.21 rushes/gm while SMU and Tech quarterbacks as a group averaged 3.3 and 2.6 receptively.

The next important stat to look at is the yards/att. Since a majority of the passes are focused within the first five yards of the line of scrimmage with the receivers being expected to get upfield from there, the yards/att average is generally between 7-8 yards. McCoy: 7.5, SMU: 7.8, Tech: 7.5.

This isn’t to say McCoy isn’t accurate, or can’t be good at some point in the NFL, or that Texas is a gimmicky offense. In the college game, coaches find what you are best at and use that as efficiently possible because if you don’t win, you get fired. Its our job as evaluators to find what these players do best in the scheme they play in and see if those skills can be used in the pro schemes run by most NFL teams. What this is meant to illustrate is just how similar the scheme Texas ran with McCoy was to those run at SMU and Texas Tech. Because a bigger statistical effect of these systems are inflated completion percentages. Its much easier to complete passes when your receiver is closer to the line of scrimmage and even easier if the team runs a lot of screens. And when the offense is additionally designed to run as a constant two minute drill to build tempo, it helps the effectiveness of these passes.

So without further ado, the completion percentage for our three examples: McCoy, 70.6; SMU, 60.6; and Tech, 67. Colt manages 10 percent more than SMU, but just 4 more than Tech. Now if we include everything we discussed on how talent levels of the players involved can effect completion percentage, its not hard to imagine that Texas would be better at running the system than Tech and SMU based on the better level of talent Texas can acquire over its fellow in state institutions.

It is also important to note that a great many of quarterbacks with high completion percentages in college go on to be huge busts in the NFL. JaMarcus Russell was a 67.8% his senior year at LSU, Brady Quinn was 62%, Matt Leinart 65.7%, and Alex Smith was 65%  at Utah. Smith is being given one last chance, but Leinart was recently cut after being a top ten pick, Quinn was traded this off season for chump change and may be Denver’s third QB, and I assume most of you reading have never heard anything nice about Russell since he made the jump to the NFL.

So lets bring everything we have so far back to our initial point. Completion percentage may not show the accuracy of a passer, but it can help in another way. Since their are so many factors involved in creating the completion percentage, we find what the common thread between the factors are. Since every factor of the passing game is involved, from the scheme, to the quarterback, to the receiver, the logical conclusion would link the stat to how efficiently the passer runs the offense. So having a high completion percentage in college is still important, but even the completion percentage can be represented in another stat, the Passer rating. Since that can be a Notebook entry in and of itself, for now I will simply reference this well written argument that delves into how passer ratings are complied at the college and pro level and how college passers rate using the NFL system.

Hopefully this has helped clarify how completion percentage is too diluted with outside forces to have a direct correlation with the accuracy of a quarterback. As is the case in scouting, the only real way to know is to watch the players play.  I don’t know about you, but I will use any reason I can to watch more football. Here in the Scout Notebook we are going to tackle more areas of how to scout each position, compare players, debate players and scouting concepts with other writers, and generally try to advance what it means to scout players as amateurs for our favorite professional teams. Questions, comments, or slanderous outbursts? Continue the conversion in the comments section and stay tuned for more features at NFL Mocks!