Wilson vs. Kaepernick: Facts & opinions on "The Edge"

Wilson vs. Kaepernick: Facts and Opinions

ESPN's analysts weighed in on which quarterback has "the edge". I weigh in on their weighing in. We'll go way into the stats looking for a way into the insights that reveal which quarterback is way into winning.

Opinions, even unsupported ones, are all part of the fun of being a football fan. So when ESPN's "The Edge" recently compared Colin Kaepernick and Russell Wilson, I have no problem with their stated preference per se.

The meat of the analysis, however, was not... meaty. Perhaps the assignment delegation went something like this:

PRODUCER: I want a comparison of Russell Wilson and Colin Kaepernick. Whatcha got?

ANALYST: Well, for starters, Kaepernick is more athletic, and Wilson is known for his intense preparation. They both--

PRODUCER: Whoa! Way too much detail! Can you boil that down to six words?

ANALYST: Um...

PRODUCER: Work on it.

Mark Brunell and Antonio Pierce compared the two quarterbacks, awarding their eponymous "Edge" to one or the other in five different categories (Arm Strength, Accuracy, Improvisation, Reading Defense, and 'To Win One Game'). At no point did the analysis go beyond what you could have guessed from a pre-draft scouting report. The whole thing had the feel of a term paper being written on nothing more than a one-paragraph entry in the encyclopedia. C'mon, guys. We've got a whole week and only two games to talk about.

Arm Strength

Brunell called this an "easy one", awarding a check mark to Kaepernick for his admittedly exceptional throwing power. Fair enough. This one, at least, can be objectively verified (someone put a radar gun on Kaepernick during a game last year; he can certainly throw it hard).

However, Wilson is hardly a weak-armed NFL quarterback squeaking by in a West Coast Offense that doesn't ask him to throw the ball more than 40 yards. Wilson has plenty of zing and can easily chuck the rock 70 yards. He can make all of the throws. Whatever additional velocity Kaepernick can impart provides no additional advantage, at least not until Kaepernick develops the ability to hit super-tight windows a' la John Elway or Brett Favre in their primes.

Jason's Edge, Arm Strength: Kaepernick, but irrelevant.

Accuracy

Brunell awards his next check mark to Wilson, with all the subtlety of a man handing out candy to children (one for you... one for you...). Again, the assessment isn't totally wrong. But it's very incomplete.

In his last game versus the Panthers, Kaepernick was 3/7 on deep balls for 88 yards. I watched every one of these passes as a refresher (I've seen plenty of his throws before) and it's clear that he the ability to hit a pinpoint pass. The deep completions to Quinton Patton and Michael Crabtree were absolute lasers. His misses downfield tell a story not of bad aim, but bad mechanics (and full credit to Brunell for pointing this out). On nearly every such errant pass, Kaepernick spotted a receiver and released the ball before getting his feet set properly. But whether this is a bad thing or not depends on context-- by getting the ball out quickly, Kaepernick takes fewer sacks; by passing deep, even with spotty accuracy, he'll get some completions and spread out the field. In other words, he's making a tactical decision to sacrifice accuracy for faster delivery.

Wilson, meanwhile, has missed a few passes of late. But he doesn't miss often. And he has a noticeable advantage in his ability to time his passes to hit a moving target.

Jason's Edge, Accuracy: If throwing at a tire, it's a tie. In an actual game, Wilson by a hair.

Improvisation

Antonio Pierce took a turn here, renamed the category "Improvisation/Running" and gave the check to Kaepernick.

Were you that pressed for time?

Let's do one at a time. Both quarterbacks excel against the blitz. Wilson, however, is far better throwing on the run. We don't have stats readily available for 2013, but the "outside the pocket" QBR from 2012 shows a massive gap between Wilson and any other listed quarterback. As further confirmation, Wilson shines on designed rollouts, where Kaepernick was just average. Even when he takes off to run, Wilson keeps his eyes downfield and is capable of releasing the ball right up to the line of scrimmage.

We know that Colin Kaepernick is faster running in a straight line, and he's broken more big runs. But, on average, the difference between the quarterbacks is surprisingly narrow. Excluding kneel-down plays, for 2013:

Wilson:
82 rushes
554 yards
1 TD
6.76 average

Kaepernick:
82 rushes
534 yards
4 TD
6.51 average

You can barely squeeze a hair between the differences in average and total production. Kaepernick has more rushing touchdowns, but these are offset by his two fumbles that occurred while running.

Wilson has 29 rushing first downs and Kaepernick 28; we can give a slight edge to Kaepernick here because 12 of his first downs came on 3rd & 4th down (as compared to just 8 for Wilson).

On the other hand, Wilson has faced, on average, a slightly tougher slate of opposing defenses to run against this year. Football Outsiders takes this into account when they calculate a rushing DVOA (defense-adjusted value over average) of +23.1% for Wilson and +11.6% for Kaepernick.

Jason's Edge, Improvisation: Wilson, because I've watched him more.
Jason's Edge, Rushing: Push. Tie. Dead Heat. Kaepernick runs faster, but he's no more effective.

Reading Defense

Sticking to the "athletic vs. studious" narrative, Pierce gives a check mark to Wilson. But this is where over-simplification reaches its peak. Quarterbacks are expected to "read" a defense before the snap and make adjustments or plans accordingly. That's one skill. After dropping back to pass, they are expected to go through their progressions (see which receiver, if any, is open); that's another skill. Then they have to make a decision to tuck it and run, scramble for time, throw the ball away, attempt a pass as designed, attempt a pass outside the play design, etc. This decision making is yet another skill. In fact, it's better described as a trait. Unlike accuracy and arm strength which can be described (if not outright measured) on a linear scale, two different quarterbacks can make very different decisions based on risk tolerance and complementary skills (e.g., running) without one necessarily being better than the other.

So in defiance of journalistic tradition, I'm going to give an honest answer here:

Jason's Edge, Reading Defense: No clue.

To Win One Game

Finally, we get to the climactic thesis, or as I like to call it, "the punch line". Both Brunell and Pierce chose Kaepernick, because "He's playing well right now," and apparently Wilson (and/or the Seahawks) are not.

Wilson's numbers are down over the last five games. San Francisco is on an eight-game winning streak. But now hold on: Why is "last five games" the predictive model in one instance and "last eight games" the predictive model in the other? Is this merely a Multiple Endpoints Fallacy?

Let's look at the two quarterbacks' recent performances. For a baseline, we'll use Net Yards per Pass Attempt, a simple statistic which measures yards/pass counting all sacks as passing plays (and sack yards as negative passing yards). Unlike ANY/A (Adjusted Net Yards per Pass Attempt), NY/A gives no bonus for touchdowns and no penalty for interceptions. ANY/A is a better descriptive stat, but NY/A is considered a better predictive stat.

For some added perspective, we'll throw an average measure of the opposition's pass defense quality, using Football Outsiders' pass defense DVOA. Note that a negative DVOA indicates a better defense.

WILSON
Last 7 games ny/a: 8.04 (average opposition pass DVOA: -1.7%)
Last 5 games: 6.10 (-4.94%)
Last 2 games: 7.2 (-2.2%)

KAEPERNICK
Last 7 games: 6.89 (-0.36%)
Last 6 games: 6.63 (-2.53%)
Last 5 games: 6.95 (+3.82%)
Last 2 games: 6.5 (+2.5%)

Kaepernick has certainly played well across any recent time frame, beating the league average 2013 ny/a of 6.21 and adding some good running performances (which aren't measured by the stat). But any difference in consistency may be illusory: Kaepernick has put up nearly the same stat line regardless of opponent, whereas Wilson's numbers generally reflect the quality of opposition defense over a given stretch.

More to the point, it takes a very specific selection criterion to reach the conclusion that Wilson is in some kind of slump. There's nothing inherently wrong with a deliberate recency bias in predictive statistics, but if such is applied based on an existing model instead of with a desired outcome in mind, Seattle's offense is doing just fine.

But wait, there's more.

Wilson's statistics (e.g., passer rating) have taken a bit of a ding thanks to three interceptions in his last five games. But those deserve a huge caveat. Against San Francisco in week 14, Wilson ended the game with a 70-yard Hail Mary to the middle of the field. That's one interception that has no predictive relevance on Wilson's ability. Against the Cardinals in week 16, the Seahawks' last possession ended with an interception that bounced off Doug Baldwin's arm and, despite the official ruling, the ground. Whether you agree with the 'ground' part or not, this clearly isn't an interception that has any predictive relevance.

Even the 'legitimate' interception against the New York Giants was the result of a jump ball on which Seattle receiver Ricardo Lockette failed to make a decent play. Fair enough to count this one against Wilson (he chose to throw), but keep in mind that it was not the result of a failed read wherein an unseen defender stepped in front of the ball.

So: Excluding the two obviously non-predictive interceptions, Wilson has committed one turnover (one interception, zero lost fumbles) in his last 32 quarters of play.

Now that's a hot streak. Or maybe a cold streak, but of the good kind. Wilson is being asked to play conservative, manage the game, and rely on Marshawn Lynch and the defense until otherwise required. Don't believe that? Look at Wilson's numbers versus the Saints when broken down by the size of Seattle's lead:

Up by 14+ points : 0/3, 0 yards, 1 sack, 0.0 any/a
Up by 7-13 points: 5/10, 82 yards, 1 sack, 7.45 any/a
Up by 0-6 points : 4/5, 21 yards, 1 sack, 3.5 any/a

How about the previous four regular season games, to round out our total of "last five" and include two Seattle losses? When trailing in the 4th quarter, this is what Russell Wilson did (excluding the Hail Mary and counting the Baldwin arm/turf bounce as an incompletion):

16/25, 231 yards, 2 TD, 0 INT
Passer Rating: 120.6
ANY/A: 8.62

In the loss to San Francisco, Wilson left the field with a 1-point lead and didn't get the ball back until there were only 21 seconds left.

In the loss to Arizona, Wilson led a 6-play, 61-yard touchdown drive to put Seattle ahead with 7:32 remaining. After Arizona retook the lead, the final comeback attempt was thwarted by a bogus/freak interception.

In the "too close" win against the Saints, the Seahawks had a 15-point lead the last time Wilson left the field.

And finally, if we're going with the "win one game" narrative, and the obvious game in question is this Sunday's NFC Championship, we can't overlook that Wilson has one other tremendous, borderline "unfair" advantage: He gets to play against the San Francisco defense, which is pretty good, whereas Colin Kaepernick has to play against Seattle's, which is great.

Jason's Edge, To Win One Game: Russell Wilson