Statistical Exposure of a Quarterback

150 150 nofriendo

A NOTE FROM THE EDITOR: This is ridiculously long. If you are on a mobile device,you may consider plugging it in so your battery does not die before you get through it. Additionally, ice can help alleviate scrolling finger cramps.

Adusoron’s Measured Critique of BKOB’s Pro-Seattle Hyperbole Disguised as “Knowledge”

Sometimes you can’t let a good troll piece go unanswered. Sometimes enough is enough, and reasoned analysis must triumph over NorthWestBubble “analysis.”

By now most of you know that our esteemed colleague BKOB (owner of the illustrious Golden Boner trophy – MSRP $.25 and presently listed for sale on eBay) is fond of allegedly “arming” his fellow hawk fans with knowledge. Usually this knowledge entails making Russell Wilson look competent at all costs, and disparaging any opposing QB at all costs.

Any person can massage or torture a set of statistical data to fit a preconceived conclusion – otherwise a whole host of NFL commentators, ESPN anchors, and denizens of the Pacific Northwest would be out of a job or hobby. This article is aimed at investigating multiple reputable advanced statistics metrics from different statistical providers with the end goal of providing a comprehensive look at each NFC West QB’s play and value through Week 6. The point of this article is not to highlight one statistic over the other, but to provide multiple lenses from which to view our favorite and/or hated rival’s QB(s).

Quarterbacks (Through Week 6)

 

SaW3WLd

Analysis:
Right now, Kaepernick is playing the best, with Palmer falling in second place. Wilson’s play has been uneven and has dropped off, Davis is an early surprise, and Stanton has been a quality backup, but not much more.

In the first chart (the standard raw statistics), Kaepernick has the most yards, passing TDs, tied for the fewest fumbles and more rushing yards. Wilson is second with the best completion percentage and the most rushing TDs. Palmer’s sample size hurts him here, so it’s hard to really compare him fairly. He has managed to avoid throwing interceptions through two games, but it remains to be seen if he can keep that up.

The advanced statistics chart tells a more interesting story. Let’s investigate further.

Traditional Passer Rating
Passer rating is a statistic with which we’re all familiar and is heavily pimped by the NFL websites. While it also has its strengths and correlates highly with winning (though not as highly as ANY/A), a significant criticism is its’ exclusion of some statistics such as rushing yards and TDs, and lack of accounting for context (i.e. a 5 yard pass that the WR takes 65 additional yards for a TD vs a 70 bomb thrown for a TD – both result in a TD, but one was the primary product of the WR, and one was the primary product of the QB), and double counting of completions. Be that as it may, Palmer, Wilson and Kaepernick are separated by 3 points between them at this point in the season and all three have good passer ratings. The issue with ending the analysis here is that the inquiry can go much deeper and distinctions can be statistically drawn between all three QBs play.

Yards Per Attempt
Yards Per Attempt (YPA) is which is an efficiency stat that demonstrates how many yards a QB will gain on average with each pass. The theory is that the more yards gained per pass, the more likely the team is to move into scoring position and therefore increase its chances of winning. Critics of this metric complain it favors Air Coryell vertical offenses and penalizes West Coast offenses. Supporters counter that it takes greater physical skill to accurately pass further downfield, so it should favor offenses and QBs who can execute those passes.

At any rate, Kaepernick has the highest yards per attempt in the NFC West. Austin Davis is a surprising second, and Stanton is the worst. Wilson is merely third, and if the Hawks continue to play like they did against the Cowboys, logic dictates Wilson’s YPA will decline. That is TBD as they continue the never-ending quest to determine if Harvin can do anything other than line of scrimmage work such as jet sweeps or bubble screens.

Adjusted Net Yards/ Average
ANY/A is a pure mathematical formula that is publicly available and quantifies the contributions of a quarterback’s passing game by including five key passing statistics; passing yards, passing touchdowns, interceptions thrown, times sacked and yards lost to being sacked. This measure is considered to be the more correlated to winning, even higher than Passer Rating. The average ANY/A will vary from one season to another but is generally around 5.35 ANY/A.

While one must make the obligatory “small sample size” qualifier, Palmer’s good start is a development worth monitoring in the coming weeks. His and Kaepernick’s ANY/A measures are terrific and are well above 5.35 average, and both the Cardinals and 49ers lead the NFC West as a bit of anecdotal support. Wilson’s ANY/A is not as good, but is above average. Davis and Stanton are closer to average, though time will tell as Davis plays out the rest of his first season starting.

Total Quarterback Rating
QBR is a polarizing topic, namely because ESPN (Eastcoast SPorts Network) is “the man” of sports media, and people resist obeying “the man.” (Power to the people and all that.) Also, there are questions about the secret proprietary formula, the influence of the clutch weight factor, and the subjectivity in assigning blame. However, those questions aside, other reputable websites such as Football Outsiders consider QBR alongside their own analytical methods so it is worthwhile as a consideration alongside the other advanced stats. QBR has heavily mitigated the clutch weight factor, and they have sought to remove the subjectivity from the statistic. The end result is a metric that accounts for a quarterback’s total effect upon the game and encompasses principles from Win Probability Added (see below). A QBR rating of 50 is average, and above 65 is pro bowl quality. QBR correlates higher than Passer Rating, ANY/A and DVOA to winning.

Palmer is surprisingly leading the division in QBR, followed by Kaepernick. Both are playing at a QBR-rated pro bowl level. Wilson is a little above average, but is only 3 points above Austin Davis. Drew Stanton, predictably, is playing at a submediocre level but the Cards managed to win around his play.

Wins Probability Added
Advanced NFL statistics Win Probability Added measures a play’s impact on the outcome of a game play based upon the expected number of points a player is supposed to generate per mathematical models of past NFL games. The theory is that if many past quarterbacks have made similar decisions within the same parameters with certain outcomes, the current QB can be measured against that model. A likely criticism of this model is that it suffers from groupthink, or assumes that the prior outcomes are the only correct outcomes, but that debate aside, the NFL is an old business and past experiences can inform on current play. Also, WPA has the highest correlation to winning of all the metrics including Passer Rating, QBR, DVOA, ANY/A, or DYAR. Application of the WPA model is where we first begin to see the most separation between the NFC West quarterbacks.

According to WPA, Kaepernick is the 12th ranked QB and first in the NFC West. Palmer is the 18th ranked QB and second in the NFC West. Wilson is the 25th ranked QB overall and the worst in the NFC West through week 6 – worse than Drew Stanton and Austin Davis. What exactly does this mean? Specifically it means that Kaepernick’s play demonstrates he is contributing more to the 49ers winning games than Palmer or Wilson are contributing to their teams’ wins.

Defense-adjusted Yards Above Replacement
Football Outsiders’ Defense-adjusted Yards Above Replacement is a measurement of a player’s total overall value, expressed in yards. Football Outsiders has calculated what an average replacement level player would produce, and then calculates the yardage above replacement level to rank QBs. For example, in 2013, Tony Romo was seventh among quarterbacks with 839 passing DYAR. Football Outsiders’ model estimated that a generic replacement-level quarterback, throwing in the same situations as Romo, would have been worth 839 fewer yards. Like with WPA, application of the DYAR model confirms a high degree of separation between the NFC West quarterbacks through week 6.

Looking at DYAR, Kaepernick is the 12th ranked overall QB and first in the NFC West. Wilson is the 27th ranked QB overall and the worst in the NFC West through week 6 – worse than Austin Davis and Drew Stanton. Again, what does this mean? It means that Kaepernick adds more total overall value, expressed in yards, over a replacement level QB, to the 49ers than Wilson adds in total overall value to the Seahawks over a replacement level QB.

Defense-adjusted Value Over Average
DVOA measures a team’s efficiency by comparing success on every single play to a league average based on situation and opponent. Player DVOA is more of the same – it is a measurement of the value of a player on a per-play basis. Football Outsiders describes their methodology and assumptions at length, but in DVOA has gained wide acceptance in the NFL world and blogs and websites such as SI.com, ESPN, NY Times Online, SB Nation, and CBS Sports utilize their data and conclusions. Like with WPA and DYAR, the DVOA model also demonstrates a high degree of separation between the NFC West quarterbacks through week 6.

Carson Palmer has the 9th best DVOA rating overall, and is first in the NFC West. Kaepernick is at 12th overall and second in the NFC West. Wilson is the 27th ranked QB overall and the worst DVOA in the NFC West through week 6. This means that Palmer is presently adding the most value to his team on a per play basis, while Kaepernick is 3 spots behind him in second place. Wilson is adding the least value to his team on a per play basis – less than Austin Davis or Drew Stanton.

The Week 6 Conclusion (So Far):
There are many weeks of football to be played, but right now, the plurality of metrics show that Kaepernick is playing the best in the NFC West, with Palmer coming in second (admittedly on a two game sample size). Wilson is up and down depending upon the statistic one chooses, but clearly the advanced statistics judge him very harshly. Again, this is why when a statistical analysis is presented; the inquiry must go deeper than selectively picking one favored statistic and eschewing all others. Also, rankings and model results will certainly ebb and flow as the season evolves, so stay tuned.

(How’s that for feedback, BKOB? Remember, at one point cavemen thought fire was best produced by rubbing sticks together. However, discovery/invention of accelerant and fire starting devices has advanced society beyond caveman status. The moral here is don’t remain stuck on caveman stats for analyzing NFL players. Embrace advanced metrics.)

Sources:
WPA   QBR   ANY/A   DYAR and DVOA    Correlation with Winning   YPA

AUTHOR

nofriendo

All stories by: nofriendo
2,344 comments

Comments are closed.