Nick Foles and Robert Griffin III: A Statistical Comparison
If NFL prospects could be reduced to just the numbers they put up in college, I think people would look at this years’s NFL draft quarterback class much differently. Luckily, we’ll never need to resort to using nothing but statistics to grade quarterback prospects, but if I were to tell you that Kellen Moore and Brandon Weeden were the two best players in this draft, wouldn’t you believe me?
You wouldn’t? Ah, well that’s okay. Today I’m only looking to compare one prospect who is under first round consideration to another prospect who is under mid-second round consideration, using a statistical perspective.
The statlines of:
|QB Name||QB Starts||QB Wins||Completion %||Attempts||Sack||Sack %|
|Robert Griffin III||41||23||67.11%||1192||79||6.22%|
|2012 QB Class Average||35.68
|QB Name||Sacks as Incomplete||Negative Play %||Consistent Play %||Sack Projection|
|Robert Griffin III||62.94%||7.55%||
|2012 QB Class Average||60.43%||7.18%||56.29%||
|QB Name||Year 1 Value||Year 2 Value||Year 3 Value||Year 4 Value||Variance||Variance/PA|
|Robert Griffin III||2091||481||3501||4293||1676.01||1.41|
|QB Name||Total Yards||YPA|
|Robert Griffin III||10366||8.7|
Lets get the important stuff out of the way first. The higher efficiency college quarterback is Griffin III, and it’s not all that close. Griffin played quarterback for Baylor, so I’m not suggesting you pretend that he’s somehow just a product of his supporting cast. But the Baylor offense was talent-loaded compared to the Arizona offense from this past season.
But what is more important is that we identify which college statistics best predict themselves at the pro level: completion percentage being the big one, but also yards per attempt, sack rate, and maybe the best predictor of overall success of highly drafted players, college games started. Also, based off of success rate concepts, I created a couple of metric for college players: 1) an adjusted completion percentage that treats all QB sacks as incomplete passes, 2) negative play percentage, which combines the likelihood of a sack with that of an interception (killing a drive), and 3) consistency percentage, which takes that negative plays and adds in incompletion rate (1-completion percentage) to see how often a quarterback completes a pass that moves the change, and penalizes for incomplete passes that are also interceptions. What consistency percentage aims to do is neutralize (as much as possible) the effect of the college players’ offense, although for quarterbacks who throw behind the line of scrimmage a lot, it necessarily will overrate them a bit because the denominator is passing attempts.
I also have plain old passing yard variance, which looks at how consistent a prospect was in college year to year in terms of total passing yards. Finally in the bright gold and green Baylor Bears colors, you can see career college YPA and an NFL rookie sack rate projection based off a linear regression for every QB picked since 2005. In all, we are analyzing the prospects by 9 different statistical passing categories, half of which were created for this comparison. At the very end, I’ll go deeper and re-examine the effect of rushing stats on QB success rates, looking back at my Ryan Mallett article from last year as a guide.
Before we even get to analyzing the numbers, what stands out the most is how similar the passing career paths of Griffin and Foles are. It’s why the variance edge flips when you divide by passing attempts. There is no advantage in how they got where they are. The only advantages as prospects from a stats perspective is who they are.
When you analyze that stats, you realize that it is still very close between the two, but a couple of clear trends emerge:
- In stats that consider and project sacks, Foles comes out ahead, sometimes significantly ahead.
- In stats that consider yards per passing attempt or vertical yards, Griffin usually comes out significantly ahead.
- When we try to separate QB stats from the context of their college offenses, Foles advantages tend to stay while Griffin’s disappear
- However, in composite categories that favor Foles, the difference is slight, sometimes very slight.
It is critical to remember that we’re not considering the impact of Griffin’s rushing ability on the college game or the pro game here, and considering the weaknesses of Foles as a runner, we cannot just ignore entirely Griffin’s speed and ball carrying ability. It is critical to find a way to work it back into what we are doing above.
You might be surprised to learn that Robert Griffin III’s career high in rushing yards (843) and rushing YPC (4.9) at Baylor came in his Freshman Season of 2008. He carried the ball 173 times that season. Before you blame his 2009 ACL tear with a hesitancy to use his physical gifts, do consider that he rushed 149 and 179 times in 2010 and his Heisman winning 2011 respectively. Griffin wasn’t running less often later in his Baylor career, nor was he running less effectively. He just happened to set his career highs in rushing four years ago as an 18 year old Freshman quarterback in the Big XII.
Foles never posted positive rushing numbers, which essentially means that his career rushing attempts (105) total was comprised more than halfway by career sacks (57) and kneels. Foles doesn’t have great athleticism, and may be destined to reprise the career of Kerry Collins in the pros: under-appreciated even if effective for 20 years.
The biggest problem here is that the quarterbacks who never ran in college (which would be a good way to define Nick Foles) have been drafted a lot lower than quarterbacks with requisite athleticism, and it’s hard to make a case that they’ve been under-drafted. I can probably make a case that New England backup QBs Ryan Mallett (3rd) and Brian Hoyer (undrafted) were picked to the Patriots because they were underdrafted. Chad Henne probably was underdrafted. But Brian Brohm certainly wasn’t underdrafted, and neither were John Beck, Jimmy Clausen, Matt Leinart, Brody Croyle, or Andrew Walter.
The results of the Ryan Mallet case study haven’t changed in the last year, if anything, they’ve gotten stronger thanks to the rookie emergence of Cam Newton and Andy Dalton, both of whom had strong college rushing numbers. But in Foles’ case, he’s a rare prospect in that as good as he is at avoiding sacks, -2.8 YPC in college is tough to do for any quarterback. Even Mallett came in at -1.1 YPC. Brian Hoyer had a -2.2 YPC figure at Michigan State. He went undrafted. Henne had a -1.8 YPC. Matt Moore had a -1.4 YPC. Kellen Moore, the scrawniest guy in the draft class, only posted a -1.3 YPC at Boise. Brodie Croyle a -1.0 YPC. Dan Orlovsky: -1.5 YPC. Tim Couch: -1.8 YPC. Jimmy Clausen -2.0 YPC. Rex Grossman -1.1 YPC. A lot of mixed results in this group of comps. Not a whole lot of return on investment. Then again: Tom Brady -1.7, Eli Manning -1.1, and Peyton Manning -1.2 YPC all turned out kind of alright.
Only a couple of other Pac-12 (nee 10) QBs, Andrew Walter (-2.7) and Derek Anderson (-2.9), offer a similar YPC average to Foles, at Arizona State and Oregon State respectively (although this puts Brandon Weeden’s -3.8 YPC at Oklahoma State into impressive context).
Foles passes the eye test in that I can say he does well throwing on the run and completing passes on the run even if he is never going to threaten the line of scrimmage as a rusher, a skill that neither Walter or Anderson had. Obviously, he’s not like Robert Griffin when he breaks the pocket. I think Foles plays well out of the pocket. But his career passing yards per attempt figure is lower in every year of his career than Robert Griffin’s career figure. If Griffin is a sensational player when he breaks the pocket, Foles is merely an average one.
There is no argument for preferring Foles over Griffin, and Griffin seems certain to justify being a high pick in the 2012 NFL Draft one way or another. Foles’ passing statistics in college — at least the ones that predict NFL performance — are right at the level of Griffin’s. He can make all the throws Griffin can from all the planes Griffin can. He probably did what he did in college with less talent around him, even though Griffin was tasked with resurrecting a program. Foles was tasked with saving a head coach’s job, and he did not succeed.
There are very few comparables for either. Sam Bradford is one, but he lacked Griffins athleticism, and played with far more talent in college than Foles. Foles’ college statistics are almost identical to those of Brian Brohm, who was an excellent draft prospect, and a terrible NFL quarterback. Foles will almost certainly have more NFL success than Brohm, for no other reason than the amount of success that Brohm had. Eli Manning or Tim Couch are probably the best college comparables for Foles, though Manning had the advantage of being a Manning, something that Foles can never be.
Foles profiles better than pretty much any recent second round prospect this side of Andy Dalton, and Griffin profiles better than any player who didn’t go first overall. The only difference between then is their legs. And history shows that this alone might decide which of them is successful (or remains healthy) and which fails (or cannot stay on the field).