Does a Quarterback's Defense Impact His Performance?
Russell Wilson is in the Super Bowl again, which means we're on the verge of hearing a lot of "his defense makes him look a lot better than he is" remarks.
I guess it's better than deflate-gate.
There's no statistical proof that a strong defense helps quarterback performance. A top-notch defense will certainly help a team win games, and that's mostly what's happened with Russell Wilson -- he's a really good quarterback, don't get me wrong, but the Seahawks' defense is a very big reason they're in the Super Bowl for a second straight season.
The issue we run into with this argument is misinterpretation and exaggeration. Some folks seem to think a quarterback is good or serviceable if his team wins, without really dissecting his actual performance. For example, Trent Dilfer won the Super Bowl with the Baltimore Ravens back in 2000, so therefore defensive performance aids quarterback play.
This really isn't the case. In 2000, Dilfer finished the regular season with a -25.11 Passing Net Expected Points (NEP) total, good for a -0.10 NEP per drop back rate. That ranked 31st of the 40 quarterbacks who dropped back to pass 200 or more times in 2000. In other words, Dilfer played far below expectation that year, and if you dig into the numbers, it continued in the playoffs. He wasn't good. His team was.
That's the misinterpretation piece. Rather than actually looking at how a quarterback performed, he's lumped in with a team metric -- wins. Just because Trent Dilfer won the Super Bowl doesn't mean there weren't 30 other quarterbacks who could have done the same thing on that Ravens team in 2000. Because there were.
Then there's the exaggeration side of things. It seems as though people can't admit a quarterback is good when his defense is strong as well. Ben Roethlisberger has fought this his entire career, despite the fact that his numbers are consistently strong. The same, as I noted earlier, is happening with Russell Wilson.
I'm here to, once and for all, crush this supposed strong relationship between a good defense and good quarterback play. Because there's really nothing that suggests a relationship exists.
A Bird's-Eye View
I alluded to it earlier, but at numberFire, we have a nifty metric we use called Net Expected Points, or NEP. In essence, NEP looks at how a player or team performs versus how they're expected to perform. If a dude gains three yards on 3rd-and-3, for example, that's a much bigger deal than a gain of three on 1st-and-10. For more on the metric -- and if you don't fully understand it, please learn more about it -- check out our glossary.
To figure out if there's been any relationship between a good defense and an efficient passing offense, from a high level, I simply looked at the correlation coefficient (r) between our schedule-adjusted pass offense and total defense NEP numbers. Because defense "takes away" expected points, good defenses have negative NEP totals. As a result, if there's anything to this relationship between passing efficiency and strong defensive play, I'd expect the r value to be negative -- the better passing efficiency (positive), the better the defense (negative).
I took every schedule-adjusted NEP data point, both Passing NEP and Defensive NEP, since 2000 and found that the correlation between the two was a measly 0.05. Keep in mind that finding a value of zero means that there's absolutely zero correlation, while 1 or -1 shows strong correlation. In this case, everything was insignificant.
Why Does This Narrative Exist?
I guess you could say that's some sort of end-all to this argument, but analyzing the data took me on a journey into digging into some individual players and to dissect the narrative a bit more.
Why is it that we assume defense helps quarterback play? Do we only use this rule for some quarterbacks?
Let's start with the first question. It seems natural to think a good defense would place a quarterback into favorable field position, and from there, the quarterback will have more opportunities to score. In addition -- and this is something people tend to bring up with Russell Wilson -- good defenses take pressure off quarterbacks, which allows them to kind of do their own thing on the football field.
I think both of these things make sense intuitively, but you have to think about the flip side of things as well.
If there's a bad defense, that means a quarterback may have more opportunity. And actually, the correlation between passing plays run and Defensive NEP since 2000 is a fairly strong 0.30, suggesting that the worse a defense is (remember, that would be a positive NEP value), the more passing plays there will be. There's no true causation here, but this does make sense.
What may not be thought through, however, is the fact that more passing plays means better quarterback performance, at least in cumulative terms.
You see, in 2014, 37 quarterbacks dropped back to pass 200 or more times. Only five of these quarterbacks had a negative per drop back NEP. In other words, nearly every starting quarterback in the NFL is enhancing his numbers when he drops back to pass, no matter his circumstance.
The reason for this is because passing is much more effective than running. The opposite is true if you were to look at high-volume running backs.
But what this could tell us is that having to "do more" as a quarterback isn't really a detriment to that quarterback's sum play. Drew Brees, for instance, had a 0.17 per drop back Passing NEP this season, good for sixth best in the NFL. Though this efficiency may dip a little with more volume (which would need to be proven), even if it did and he threw 200 more passes in 2014, he still would have been adding to his production with each drop back.
In this case, a good defense may actually hurt a quarterback's numbers when looking at things from a cumulative perspective.
The other question -- Do we only use this rule for some quarterbacks? -- is one that's a little tougher to prove. I think it's safe to say, though, that we do tend to use this defense excuse for only certain quarterbacks, and those signal-callers are typically ones in the spotlight.
Russell Wilson is worse than his numbers suggest because he has a good defense, while Andrew Luck's should be better. That's generally the narrative. Why do we target these two players specifically? Because they're big names who are currently revolutionizing the NFL.
Meanwhile, each season we see quarterbacks perform at their best level with their worst defense, and vice versa. Ben Roethlisberger, for instance, saw his best season in 2014 and, analytically, this year's Steelers' defense was by far the worst he's ever played on.
Why aren't we talking about that?
Or how about the fact that Jay Cutler's best year, which came in 2008, was paired with a defense that ranked 31st in the league according to our metrics. Instead of pinpointing this, Cutler backers are saying he performed poorly this past season because Chicago's defense was so bad. Come again?
I could play this game all day long. Daunte Culpepper was ridiculous when his defenses were bad, and he played like you-know-what when they were good. Jake Delhomme wasn't at his best when his defense was good -- his top season came in 2004, the only year the Panthers had a positive Defensive NEP total.
Don't get me wrong, it goes the other way, too. Matthew Stafford's two best seasons were 2011 and 2014. It just so happens that he played with his best defenses during those years. Matt Hasselbeck, Matt Ryan, Jason Campbell and Matt Cassel have also seen their best seasons come when their defenses were top-notch.
The point is that there's absolutely nothing that suggests a quarterback's performance is enhanced when his defense plays at a high level. For every Russell Wilson example you might give, I can find just as many who produce in the opposite direction.
A defense doesn't help a quarterback, and if it does, it's not significant at all. A running game, good wide receivers, a good scheme, coaching, the individual quarterback's mechanics -- that's what helps a quarterback perform.
A defense just helps a quarterback win.