Signature Stats: Accuracy Percentage

| October 10, 2013

2013-WK05-ACCPCTCompletion percentage is a great statistic for evaluating quarterbacks. Quite simply it’s tough to move the ball consistently through the air without actually completing passes frequently. The only problem with completion percentage is that there are so many factors that can sway the stat that quarterbacks have little to no control over.

Drops, batted passes, spikes, throwaways, and passes where the quarterback was hit as he was throwing all negatively impact completion percentage, but are also all highly random events. So while you’re getting a cut and dried relationship between completions and attempts, you’re not getting an accurate measure of performance.

To solve this, PFF offers Accuracy Percentage. The formula is as follows:

Accuracy% = (Completions + Drops) / (Attempts – TA – BP – Spikes – HAT)

Accuracy Percentage is basically how accurate a quarterback is on passes that target and reach a receiver. Now that you know how it works let’s check out the top performers.

Rk
Name
Team
Accuracy %
Completion %
Difference
1Peyton ManningDEN85.6%75.8%9.8%
2Philip RiversSD82.3%73.7%8.6%
3Matt RyanATL80.0%69.3%10.7%
4Drew BreesNO78.4%69.7%8.7%
5Tony RomoDAL77.9%71.8%6.1%
6Aaron RodgersGB77.5%66.4%11.1%
7Ryan TannehillMIA77.1%62.6%14.5%
8Robert Griffin IIIWAS75.9%62.4%13.5%
9Matthew StaffordDET75.1%63.8%11.3%
10Terrelle PryorOAK74.7%68.3%6.4%

Peyton Manning has truly been on another level this season. His Accuracy Percentage would be the best single-season mark by over four percentage points and only seven quarterbacks all season have had a single game with an Acc% as high as his is for the season. Right on his heels, though, has been the resurgent Philip Rivers. One of the main reasons Rivers’ has been much more accurate is that he’s finding a lot more receivers in rhythm. His average time to throw has dropped from 2.79 (10th-slowest) last season to 2.41 (third-fastest) this season and in turn he’s had far fewer throwaways. In 2012 he threw away one out of every 12.5 passes while this season that number has almost quadrupled to one out of 47.5 this season.

Rk
Name
Team
Accuracy %
Completion %
Difference
25Tom BradyNE68.1%56.6%11.5%
26Carson PalmerARZ67.5%58.9%8.6%
27E.J. ManuelBUF67.4%56.7%10.7%
28Colin KaepernickSF66.7%56.1%10.6%
29Joe FlaccoBLT66.0%57.7%8.3%
30Michael VickPHI65.8%53.8%12.0%
31Russell WilsonSEA65.8%58.3%7.5%
32Eli ManningNYG63.9%53.7%10.2%
33Blaine GabbertJAX62.8%48.8%14.0%
34Josh FreemanTB58.2%45.7%12.5%

What happened to Russell Wilson? Last season his 77.1 Acc% bested the likes of Matt Ryan, Drew Brees, and Tom Brady while this year he’s fallen to the fourth-worst percentage. Thankfully for them, Seattle is still 4-1 without their quarterbacks ‘A’ game. It is worrisome, though, that against the three best teams Seattle has faced (49ers, Texans, Colts) Wilson didn’t crack 60% Accuracy in any of them. Similar concerns apply to Colin Kaepernick whose Acc% has dropped 9.3 percentage points from last year. Kaepernick’s accuracy struggles have been especially apparent in his deep passing. His league-leading 60.6 Acc% on throws over 20 yards in 2012 has dropped to a pedestrian 30% this season.

Notables

•  The largest difference between Acc% and Comp% comes from Brandon Weeden. His 72.9 Acc% is 18.4 percentage points higher than his 54.5 Comp%.

•  Tony Romo, on the other hand, has the smallest difference, thanks in large part to the Cowboys having the second-lowest drop rate.

•  E.J. Manuel has found the regular season to be quite a bit tougher sledding than the preseason. His 90.0 Acc% in two preseason games has fallen dramatically to 67.4 in the regular season.

•  While there was no denying that Josh Freeman was disastrous in Tampa this season, he didn’t get much help. Freeman had 19% of his catchable passes dropped, easily the highest percentage in the league.

•  It appears as though Marc Trestman has been able to bring out the accuracy in Jay Cutler once again. Cutler’s Acc% has been below 71 every season in Chicago after he had an Acc% of 73.3 his last season in Denver. This year he sits right outside the Top 10 at 74.6.

 

Follow Mike on Twitter: @PFF_MikeRenner

 

  • a57se

    Not very good as you count a pass that was off target as an accurate pass because it was completed. A perfect example is Joe Flaccos SB TD to Jacoby jones…..Flacco threw the ball ten yards short but Jones was SO wide open he was able to stop, have a smoke, then catch the ball and still run into the end zone before the safeties got there…….that SHOULD NOT count as an accurate pass.

    • ItsJustWerner

      If the pass gets to it’s intended target, how is it not accurate lol. It could be tipped by the whole defense so long as it get’s to your own guy for all I care for accuracy.

      • a57se

        Well if you are evaluating a QB it makes a big difference as a lot of the passes that aren’t placed properly get picked off in the NFL. If a QB has really good placement on 30% of his passes vs. another QB with a 50% placement percentage…chances are the QB with the higher placement percentage throws less picks…….they may complete similar percentages of passes.

      • Lelouch vi Britannia

        Look, you should underthrow your receivers… a little bit. But there’s a point at which it just becomes ridiculous. In the ’11 AFC Championship Game (’11 season, ’12 year), Flacco pulled the same stunt with Torrey Smith. 45 yard pass to a wide-open receiver, underthrown badly, the average QB completes that throw and gives him room not to be tackled. So really he cost the team 30 yards and a score.

  • Steve Palazzolo

    Our grading system takes care of that, everything is accounted for. This is just a stat that takes into account whether or not a pass was able to be completed, not necessarily ball location.
    In our grading, there are plenty of instances where completed passes get neutral or even negative grades and conversely, incomplete passes that are thrown in good spots that could potentially earn positive grades.

  • John

    This is a good stat for removing first order biases from the raw accuracy numbers. You should, however, adjust for other factors, most notably the distance the average pass travels in the air. It is of course easier to hit a receiver in the numbers from 5 yards than 50. To the extent a QB consistently throws longer passes, his accuracy should, on average, suffer. A simple regression between accuracy and average yards in the air per pass would provide an adjustment factor for a further adjusted accuracy metric. My back of the envelope shows Cutler moving up 3 spots, Romo and Stafford falling 3 spots each, and Dalton and Bradford falling 6 and 7 spots in the rankings, respectively (not surprising when you think about their offenses). Lower down in the list Alex Smith drops 6 spots, which further validates the revised metric. Big gainers include Vick, Geno Smith and Weeden (to my surprise).