There has been a thought stuck in my head for months. This thought involves how quarterbacks are evaluated. Sure there’s QB rating, completion percentage, conversion rates, yada yada yada, and they are all well and good. But none of them focused on something I, as a fan and someone with (very, very, extremely minor) quarterbacking experience, find to be an extremely important factor in a quarterback’s play: time in the pocket and how they use it. With that in mind I set about trying to create a new statistic to measure quarterback efficiency, cleverly naming it Quarterback Efficiency Rating (QER). Now, I created a series of formulas to calculate different game situations that would assign a rating to every play where a quarterback dropped back with the intention of passing, but quickly found that attempting to time a football game with any semblance of accuracy without a DVR was impossible, so field testing couldn’t be done until I was home for break. Since coming home a few weeks ago I have tried the statistic on a very limited basis (6 total quarters of NFL football). So far it shows promise, but I know its not perfect by any stretch. There are certain situations that seem too lopsided one way or the other, certain factors in the scoring that are honestly random numbers I decided to assign certain events, and the formatting for the totals may be off-putting to some people.
With all those thoughts in mind I figured what better way to fix any problems than to let people smarter than me take a look at an critique my work. Some quick notes before a full explanation of the statistic: all yardage numbers used are not official, I used the announcers and some quick eyeballing for most of them, so they may be a little off, and the time in the pocket is all hand-timed and therefore subject to human error, but for the most part yards and time should be pretty close to the actual numbers.
Key:
QER=Quarterback Efficiency Rating
TIP=Time in the Pocket (defined as snap of the ball to release/sack/crossing the line of scrimmage)
C=First Down Conversion
TD=Touchdown
F=Fumble
FL=Fumble Lost
INT=Interception
INC=Incompletion
YDS=Yards gained/lost
Formula:
Completed Passes/Scrambles for Positive Yards= (YDS/TIP)/10+1(C)+10(TD)
Incomplete/Intercepted Passes= (TIP x -1)/10-10(INT)
Sacks= (((TIP x YDS LOST)/10)/2)-1-3(F)-7(FL)¹
Plays that result in negative yardage= (TIP x YDS LOST)/10
¹ Lost fumble counts as both F and FL on such plays, for a total of -10.
Now, this formula results largely in numbers in the range of 1.5 and lower, rounded to the second decimal place, which initially makes it hard to discern good play from bad due to the small size of the totals produced. However, when applied to a full game, and even one half, it has proven to be quite effective in indicating the superior quarterback (albeit in games where that could be easily discerned by the naked eye). The results from the admittedly miniscule sample size below.
Game 1-Colts at Jaguars, 12/17/09, second half only.
I forgot to record the full game, missing my chance to evaluate the first half. My bad.
At halftime the score was 21-17 in favor of the Colts, who would go on to win 35-31 for their 14th and final win of the season.
Both quarterbacks played well in the second half, each totaling over 100 yards and 2 scores through the air. Watching the game however it appeared that Peyon Manning was performing on a higher level than Jacksonville’s David Garrard (big shocker there), and that was reflected quite well by the QER.
P. Manning’s totals:
Passing plays: 18
Completions: 11
Passes intercepted: 0
Fumbles: 0
Lost: 0
Sacked: 0
First downs converted: 7
Total TIP: 44.11 seconds
Total Yards: 180
Total QER: 23.13
QER/Play: 1.29
TIP/Play: 2.45 seconds
D. Garrard’s totals:
Passing plays: 25
Completions: 11
Passes intercepted: 0
Fumbles: 0
Lost: 0
Sacked: 1
First downs converted: 6
Total TIP: 64.46 seconds
Total yards: 117 (includes yards lost due to sack)
Total QER: 15.55
QER/Play: .62
TIP/Play: 2.59 seconds
Using the new QER metric indicates Peyton Manning was over twice as efficient in the pocket as Garrard in the second half, and one would be hard pressed to argue that having watched the actual performances of both players. One half of football, however, is not a reliable way to verify the accuracy of the metric.
Game 2-Giants at Redskins, 12/21/09
This game is already infamous for the incredible ineptitude shown by everyone associated with the Washington Redskins, and also involves the other Manning brother.
E. Manning’s totals:
Passing plays: 29
Completions: 19
Passes intercepted: 0
Fumbles: 0
Lost: 0
Sacked: 2
First downs converted: 10
Total TIP: 74.48 seconds
Total Yards: 255
Total QER: 43.55
QER/Play: 1.5
TIP/Play: 2.5 seconds
Jason Campbell’s totals:
Passing plays: 35
Completions: 15
Passes intercepted: 2
Fumbles: 1
Lost: 0
Sacked: 5
First downs converted: 10
Total TIP: 100.32 seconds
Total Yards: 180
Total QER: -3.67
QER/Play: -.1
TIP/Play: 2.87 seconds
Again, one can’t help but say the metric proved true, as Campbell was horrendous and Manning played one of his better games all season, once again leaving yours truly with feelings of both confidence and trepidation. Since then I haven’t gotten around to testing the stat due to the holidays and preparing for next semester. That being said I have been thinking of practical applications for the statistic and the primary one seems to be in terms of scouting college and free agent QBs and determining which ones thrive given a certain amount of time in the pocket, useful for systems that require quicker decisions and teams with poor O-lines, as well as for determining which of two quarterbacks will fit best on a specific offense. I’ll probably record the BCS title game tonight and apply the staistic by next week to see how well that theory holds up.
Well, that’s it. Now it’s your turn. What works? What doesn’t? What have I forgotten or missed? Am I a complete moron wasting hvs time?
With your help I hope to answer all these questions and more.
Oh no! You’re turning football into baseball! AHHHHHHH
Seriously though, this would be a better alternative to QB Ratings. If the networks counted time in the pocket on their box scores, this would be much quicker to calculate and more accessible for people.