Christian Ponder does not have a future as a quarterback with the Minnesota Vikings.
More precisely, the likelihood that he's a successful quarterback, at least with us, is a vanishingly small prospect. Human performance, the NFL and quarterback development more specifically is a black box, and very difficult to predict, but over the past few days I've taken a look at the history of rookie and sophomore passers, and the results aren't too encouraging.
The full article is over at the Bleacher Report, but I'll quote it liberally here and add a few things here that I didn't write up over there.
I've been away once again for quite some time, now because of computer issues. I have, however invested in a Lenovo Y500 Ideapad (having arrived a couple days ago) and have thrown my 4-year-old Dell Studio 1737 to the wayside. We had good times, Dell. But now I have 4 times the RAM and twice the processing power, so goodbye. I've moved on.
That segues pretty awkwardly into the rest of this piece, where I'll argue that the Vikings should be similarly move on from Christian Ponder in favor of any number of alternatives.
The history of quarterback development is a difficult one to study. There are far too many confounding variables, and research needs to be done on relative team strength, media perception (which can speak to the positives many see from Andrew Luck despite poor numbers and why there was legitimate suspicion about Marc Bulger's success despite his league-high adjusted yards per attempt and second highest passer rating his rookie year), and a quarterback starting in his fourth year under contract is much different than one starting out of the gate.
Even good first year performances can be misleading. Yes, Cam Newton seems to be suffering from a sophomore slump (although I'm not really in that boat—that team could be like the 2011 Vikings... far better than their record) but an even more egregious example is Vince Young, who according to Norm Chow ("considered one of the finest teachers of quarterbacks in the business," according to this aggrandizing ESPN piece) clearly had "it":
"Besides the physical skills, it's the will and the strength inside, the 'it' factor that a quarterback either has or doesn't have," Chow said. "Obviously, he has it. The quickness in which he won the team over and the quickness in which he became the leader, the guy everybody looked up to, that was special.
"Vince Young has this great ability to lift others around him."
Naturally his offensive coordinator would talk him up, but this wasn't an uncommon opinion. Statistics websites like Cold, Hard, Football Facts ignored some of their commitment to raw data and called him the "real deal."
His intangibles were strong enough to overcome any technical deficiencies he encountered at the quarterback position.
Most NFL quarterbacks must learn how to play the position before they can win games. Young proved he could win while learning. He's going to be special.
There are any number of sources that agreed that he had the intangible qualities of a leader, including personnel executives from other teams and opposing coaches. Of course, the 2006 Offensive Rookie of the Year didn't pan out.
Terry Bradshaw did not have "it" his rookie year.
I only bring this up to stress how difficult it is to predict how well quarterbacks will do, and how media analysis might not be useful. Moreover, even the steady hand of statistics—which generally have a better track record of predicting success than post-rookie year scouting reports in my experience—has made some glaring mistakes. Charlie Batch lit it up his rookie year, finishing with a quarterback rating well above the league's average, somewhat rare for a rookie. He had the highest passer rating of any quarterback below the age of 28, at the age of 24.
Indeed, there were people pointing to this, calling him a better pick than Peyton Manning, drafted that same year.
Still, there are some powerful indicators of success that people can look to in a quarterback's first two seasons that might tell people what to expect. They aren't foolproof, but they can give context.
I put together a massive spreadsheet of quarterbacks who threw 160 pass attempts in their rookie and sophomore years, which excludes gems like Tom Brady and Drew Brees, but still comes up with 89 qualifying passers. The important thing to remember is that statistics need to be era-adjusted.
Jeff George threw a 73.8 passer rating his rookie year, which would make him the 32nd ranked quarterback in 2012. In 1990, his rating was nearly average. Nine years later, Tim Couch's passer rating was worse, but he still threw an average passer rating of 73.2. Nine years after that, Joe Flacco had to throw with a passer rating of 80.3 to meet league average.
What I've used to adjust these figures is to take advantage of what Pro-Football-Reference has done, which is to borrow from baseball's +metrics. That is, they assign players a percentage relative to league average. A passer rating+ of 120 would be twenty percent above the average rating of qualifying passers, and a passer rating+ of 63 would make you John Skelton, 37 percent below the average rating.
There, we can see that Johnny Unitas' 74.0 passer rating his rookie year was very impressive; it was 16 percent higher than his colleagues. By comparison, Colt McCoy's 74.5 could be a bit disappointing, as it was 89 percent of the average passer rating.
In general, rookies averaged out to 90 percent of league average in passer rating, 92 percent in adjusted yards per attempt (which provide penalties to interceptions and bonuses to touchdowns) and 88 percent in adjusted net yards per attempt (which, in addition to interceptions and touchdowns, take into account sack yards). These averages hold true throughout the entire dataset and when reducing it to the past decade. For both the entire population and the last-decade sample, the medians were similar.
As a rookie, Ponder's passer rating was 81 percent of the league average, and his adjusted yards per attempt (which I'll just refer to as AY/A) was the same. He threw touchdowns at a slightly higher rate than colleagues, but interceptions at a much, much higher rate, hurting his team while behind center much more.
That alone is not a death knell. 24 of the 89 quarterbacks finished below him in AY/A, 4 of them Hall of Famers. 3 more of them took at least one trip to the Pro Bowl. Still, that leaves 17 other passers who performed worse than him and didn't pan out to really be franchise quarterbacks.
It's not fair to compare him exclusively to those who did worse than him, so the quarterbacks who finished within two percentile points of him in either direction were compared. Of those 13 quarterbacks, only two went to the Hall of Fame (Dan Fouts and John Elway) and one more went to the Pro Bowl (Kerry Collins, whose inclusion to the Pro Bowl I don't really understand).
The other ten are a whose who of mediocrity, including Mark Sanchez.
Several other quarterbacks on that list show up on Pro-Football-Reference's worst quarterback of all time, including Rick Mirer ("winning" first place). In fourth place is Dan Pastorini and Steve Walsh falls tenth.
The contest measures total damage below replacement level player that these quarterbacks caused, so Mirer's eight seasons and Pastorini's twelve seasons counted for more than Ryan Leaf's three (although Leaf still ended up in the "top" ten). Also on the list is Kerry Collins, curiously enough. So, of the thirteen quarterbacks with a similar rookie season to Ponder, three of them ended up hurting their teams more than any other quarterbacks in history and another one ended up hurting his teams while also managing to make undeserved Pro Bowls (Kerry Collins).
One more quarterback showed up on the list as the 37th most damaging quarterback, and if the list had been created this year Mark Sanchez would have showed up in 19th place (if the season ended now).
There seems to be a certain proclivity of those who started out just shy of the rookie average for passer rating—they stay with teams too long and hurt them in the long run. Obviously, the Vikings would do best to avoid that.
Still, the possibility of improvement can't be ignored, and the small sample of similarity in one statistic isn't that useful. In rookie passer rating, Ponder matches four Pro Bowlers and no Hall of Famers while also mirroring two on the dreaded "worst quarterbacks" list. There were three other duds.
After adjusting for sack yardage, his rookie year matched Hall of Famer John Elway, and another Pro Bowler. The other six quarterbacks included four list members, and he performed worse in this category than Rick Mirer and Dan Pastorini.
There are any number of era-adjusted statistics to pore through, and it's just as easy to do it for his second year.
After tracking his improvement through two years, his base statistics closely resemble a number of quarterbacks, which I detailed in my article:
Of all the passers who posted a similar progression rate to Ponder, only one ever went to the Pro Bowl (Neil Lomax), while four others were failures. Of the nine passers that posted similar overall numbers to his sophomore year, two went to the Hall of Fame, and an additional two were participants in at least two Pro Bowls. Only Bill Munson progressed at the same rate and posted similar numbers throughout his early career, and he wasn't too impressive.
The two Hall of Famers that matched Christian Ponder don't really have much more in common with him than base statistics. A 1970 scouting report of Terry Bradshaw could sound familiar to Vikings fans. He "couldn't read defenses" and scrambled too early out of the pocket.
But the similarities end there. Bradshaw slung the ball too often as a rookie, throwing six touchdowns to 24 interceptions. His improvement came from becoming more conservative, slightly decreasing his raw yards per attempt and massively increasing his adjusted net yards per attempt. He accomplished the latter by decreasing his interceptions in a big way-from throwing interceptions nearly three times as often as average to throwing at a league average rate.
Ponder's improvement in completion rate came from an inverse tendency—one that led him to be much more conservative than before. This difference might be a stark indicator that the Florida State alum isn't really progressing in a way that points to real improvement.
Troy Aikman is much more similar, having increased his completion rate and slightly decreasing his touchdown rate. He finished with a nearly identical era-adjusted net yards per attempt and a worse adjusted passer rating—one that would grade out to 73 this year, just above Brandon Weeden's rating.
The difference here is that Aikman progressed better than Ponder, matching Ponder's impressive improvement in interception rate while also increasing his yards per attempt; he notably did not increase his completion rate by decreasing his average yards per completion, unlike Ponder whose passing yardage decreased by a whopping two yards per reception.
Overall, I found that quarterbacks with similar overall rookie years ended up being big successes about one third of the time. Unfortunately, his sophomore campaign revealed a success rate of about one in eight.
Given that ESPN Insider's own study, using Defense-Adjusted Yards Above Replacement (as developed by Football Outsiders) found a more depressing result, it could be time to move on.
The DYAR metric is basically a measure of how many yards a particular quarterback provides his team above a replacement-level player (a term borrowed again from baseball statistics, and often means about 75% of the average player). Players who posted two consecutive seasons of negative DYAR in their first two years were almost uniformly failures.
Throughout this entire process, Kyle Boller matched Ponder most closely, both in terms of progression as a playcaller and sophomore stats. The best next match might be Jeff George, although Troy Aikman (discussed above) also happens to match in some ways.
The biggest problem is that Ponder's change in yards per attempt relative to his sophomore year (also adjusted for each year) is negative 9.5 percent, pretty abysmal. The average sophomore quarterback improved his yards per attempt by 5.2 percent, and only 22 percent of quarterbacks did worse than Ponder from year one to year two.
Of those seventeen quarterbacks to decrease their passing efficiency, eleven of those passers had a rookie year above league average in yards per attempt. A high proportion of them (four: Y.A. Tittle, Bob Waterfield, Matt Ryan and Marc Bulger) performed well above league average.
The six quarterbacks who had a lower than average efficiency their rookie year, then got worse, were not that great. We see the return of Dan Pastorini and only one Pro Bowl player—Phil Simms.
Again, this is a comparison that only compares him to worse quarterbacks, so it's not complete. There were 19 quarterbacks who had similar devolutions in their yards per attempt, and thirteen threw under league average in their first year in the league. Within that set of thirteen, one went to the Hall of Fame: Terry Bradshaw, who was discussed above.
Bradshaw decreased his yards per attempt by 7.4 percent while increasing his touchdown rate by 10.8 percent, increasing his completion rate by a ridiculous 84.7 percent and decreasing his interception rate by an 84.6 percent. That increase in his completion rate was the highest in the population and the decrease in the interception rate was the third largest.
Only one other quarterback in the group went to the Pro Bowl (Archie Manning), while the eleven others floundered. Only ten quarterbacks decreased their yards per attempt and were successful (of 26) and 28 of 61 of the other quarterbacks were successful.
Only three Hall of Famers decreased their yards per attempt after transitioning to their sophomore year, and only one of them (Bradshaw) had a worse than average season in their rookie year. The other nine Canton inductees improved in their second year.
It would be easy continuing providing reams of statistics. But, I've linked the document that has all of them above, and I'll link it here, too. It's hard to read, and the original is a much easier to read Excel document that has conditional shading based on how much better or worse than average the statistic is.
The point remains the same. Ponder doesn't have the statistical backing that would have you believe he has a future in the NFL.
I find it hard to believe that the 40 quarterbacks in the 89 quarterback sample who made it to at least one Pro Bowl (or rather, 39 and Kerry Collins...) all uniformly had great receivers, nor do I believe that they all had great offensive coordinators.
There are a number of these players who had poor starts for whatever reason, but many of them had the subjective qualities people recognized that would make them great, Troy Aikman in particular. Ponder does not share that same love from personnel men or sportswriters.
Five options are available to the Vikings. Naturally, they could give Ponder another year (and they seem to be indicating they've attached themselves to that choice) while surrounding him with talent. Ponder's chances of turning it around are not zero—he has more than a shadow of a possibility of turning out to be a successful quarterback. But the odds are still long, and the Vikings would at least get their answer about Ponder's improvement curve.
The second option is Joe Webb. The rise of running quarterbacks and offenses specifically designed to exploit their talents has given much more legitimacy to Webb's skillset, but I'm ambivalent. Here's what I wrote (and I'll admit that the subtitle for this piece was a bit misleading):
He's hardly had enough time to put together an on-field resumé that screams "NFL passer": Webb has recorded only 152 pass attempts (and 41 runs). [I]n those limited attempts, he's not been great
On only 40.7 percent of dropbacks did Webb produce a successful play ... Quarterbacks have typically produced successes on over 48 percent of plays, with the bottom-level passers struggling around 43 ...
His passing has been subpar, even though the average depth of his passing targets has been much higher. With a total adjusted net yards per attempt of 3.8, Webb has produced numbers significantly worse than Ponder in every season he's attempted passes, and performs ... [at] only 76 percent the rate of other NFL signal-callers.
[O]ther running quarterbacks have had a higher pass yards per attempt, making their success more sustainable. Colin Kaepernick, who has had a similarly low number of attempts, has an adjusted net pass yards per attempt of 7.3. Robert Griffin III's ANY/A is 7.6, and Cam Newton's is 6.6. Webb's 3.8 seems pitiful in comparison.
Even if you add Webb's running yards ... his adjusted yards per play comes out to 5.8 ...
His mechanical and subjective problems vastly outpace Ponder's, and Webb's rhythm, timing, footwork and vision have all been called into question. Webb has never been immersed in an offense, which certainly should grant him some latitude, but other successful passers (including Nick Foles and Colin Kaepernick this year) have overcome that limitation.
The real question, however, is not whether or not Webb will be a successful quarterback; it's whether or not the likelihood of Webb becoming a more successful quarterback than Ponder. Ponder's high volume of pass attempts decreases most of the uncertainty surrounding his ability, and Webb's low usage conversely creates a cloud of mystery surrounding his total capability.
This is a difficult process, fraught with uncertainty and hidden information—coaches and talent evaluators with the team always know more than the average fan—and the question of whether or not the Vikings should abandon Ponder will take more than an hour to decide.
Naturally if the Vikings decide to go with Webb, they should design an offense to his strengths and avoid his weaknesses. If the Vikings fire Musgrave, they could pursue Andy Reid or Norv Turner as offensive coordinators, both of whom have keen offensive minds that could find ways to make sure Webb's athleticism, deep ball and poise are central features of the offense, while finding ways to hide his issues throwing in the middle of the field or with intermediate routes—especially with players like Percy Harvin and Kyle Rudolph, who can more than overcome such deficiencies (and have, with Ponder).
They wouldn't need to run the spread option. Reid could easily incorporate the University of Alabama-Birmingham's multiple offense concepts into his West Coast System, and Norv Turner could do it even better in his Air Coryell. The multiple offense is simple, but Webb thrived in it, and there are no real constraints on its complexity.
Players like Rhett Ellison and Percy Harvin are perfect fits, and they could adopt some of the tricks that make the Patriots so effective.
The third option is naturally an first round pick in the draft. Normally, this would make a lot of sense, but with a weak draft class and a projected pick not suited for picking a top-tier passer, I expect that this year's draft will not follow the usual trend, where half of all 1st round picks are successful enough to play at least 5 years as a starter.
If the Vikings select at 18, 19 or 20, they'll have the 50th pick in the second round (or later), so a viable quarterback prospect would be in the first round. Naturally the cost of this pick would be a top prospect in a stronger class, say defensive tackle, receiver or even linebacker. The Vikings have an equal need in those areas and would sacrifices a stronger prospect in a better class if they were to get a quarterback, especially if they were to reach.
They could of course exercise a fourth choice and pick a quarterback later than the 50th pick. For every Schaub, Brady or Hasselbeck, there are dozens upon dozens of quarterbacks picked in later rounds who amount to nothing. It's a strategy that yields a solid starter around six percent of the time. More likely to make an impact are 4th and 5th round skill players, who are much more often see the field than late-round quarterbacks.
In either of those cases, the probability of a successful quarterback could be higher than the chance that Ponder or Webb make an impact as a starter for the Vikings—Ponder's chances are very low and there are good arguments that Aaron Murray or Ryan Nassib could be better than either of them right now.
Picking a quarterback in a late round would result in a lower payoff, and there's not much of a consolation if that late-round passer doesn't beat out Ponder or Webb. While most players that don't hit it big right away can wait it out and improve over the course of the season by seeing limited time on the field, quarterbacks are different beast entirely. They don't rotate in on reps or see meaningful field action, and are much more of a zero-sum opportunity.
On the other hand, the cost is lower. The Vikings would be sacrificing a developmental defensive line or cornerback prospect, or potentially even another receiver. It might take some time to get up to speed, but the Vikings will probably know who they have and what they can do if they carry a roster of Christian Ponder, Joe Webb and a late-round prospect as they enter the 2014 draft; one with potentially better QB picks.
An interesting bit about late round prospects that applies to Joe Webb: every late-round prospect logged that ended up starting for more than five years had at least 5.5 yards of adjusted pass yards per attempt in their first 100 passes, except Matt Hasselbeck (who exceeded that number after hitting 150 pass attempts). Joe Webb is not in a spot to do that, unless you count his runs as pass attempts.
The final option is naturally to employ a quarterback that has already been playing in the National Football League. This could either mean pursuing a quarterback in free agency or trading for one. The only recognizable names who have contracts set to expire are Joe Flacco (who almost assuredly will re-sign with the Ravens), Jason Campbell, Charlie Batch, Byron Leftwich, Rex Grossman, Tarvaris Jackson and Matt Moore. Of those, aside from Flacco of course, Moore is the most intriguing.
Last year, Matt Moore ranked 12-15th in a number of passing measures, including yards per attempt, adjusted net yards per attempt, and passer rating. Certainly better than what the Vikings could expect from Ponder. If they want to rely on the run game with a passer merely to keep defenses honest, Moore provides much more than Christian. It remains to be seen if he's a better player than Webb. While I doubt it, my doubts count for little and Webb could easily outperform Moore and Ponder so long as the offense is designed to take advantage of his strengths and skills.
Rumors of Alex Smith and Tony Romo's ending tenures at their respective teams persist, and Matt Flynn is probably paid far too much for a backup (Cap hit of 4M this year, 7.25M next year and 8.25M in the final year of his contract). None of those options are necessarily appealing to me, first because I think the picks the Vikings would have to give up trading for these quarterbacks would be more valuable than these passers actually are. Alex Smith will cost 9.5M this year, 10M next year and 9M in his final year in cap hits.
Naturally Tony Romo costs quite a bit more, although he has proven positive playmaking ability and a preternatural ability to improvise. There are some prices where Tony Romo might be worth the trade, but his 16.8M cap hit is daunting. Given Matt Moore's low negotiating leverage, the fact that he wouldn't need to be grabbed in a trade, and that I genuinely believe he's better than Flynn and Smith would have me choose that avenue more than anything else.
How would you solve the Ponder Problem?