Saturday, September 17, 2011

Revisiting the Rule of 26-27-60

Back before the ’10 season, SI writer John P. Lopez published a piece examining whether there was a magic bullet that could allow teams to predict NFL success at the quarterback position.

Could a simple formula have warned us of Russell's lack of NFL readiness? And Ryan Leaf's and David Carr's and other failed, high-pick quarterbacks?

Call it the Rule of 26-27-60.

Here is the gist of it: If an NFL prospect scores at least a 26 on the Wonderlic test, starts at least 27 games in his college career and completes at least 60 percent of his passes, there's a good chance he will succeed at the NFL level.


The article gets mentioned on occasion around the sports community, and every time it does, I get irritated by it. The problem is that Lopez falls into the trap of cherry-picking his quarterbacks, and not particularly doing the greatest job at it. As such, I’d like to take a much deeper look at this.

So my intention is to look at the following: I’ll be examining all quarterbacks with more than two years as the primary starter in the NFL. This will eliminate guys like Sam Bradford and Tim Tebow, who we can’t really say whether they’ll be good or not. I do intend to include guys like Matt Stafford, because even though he’s started only 14 games in the NFL, I do think we’ve seen enough of him to say with confidence that if he can stay healthy, he’s going to be a very good NFL QB.

I’m also going to include all first round QB draft picks between years 2002 and 2009 if they’re not primary starters, as this mostly labels them busts. Given the predictor of success, we’ll want to know the QBs that fizzle out as well. Starting with 2002 is a bit arbitrary, but ’01 only had Vick who’s included in current starters and ’00 had Pennington who is difficult to categorize anyway, so it’s not terrible not to include him.

I’m also rounding the study out with two other QBs that deserve mention here, Peyton Manning and Brett Favre. Both are long-time starters that just finished or are injured and would normally still be starting. This gives us a population of 40 different quarterbacks, and should be a good gauge of whether the Rule really does work or not.

I’m also going to try to simplify the definition of QB success by listing four different categories. They’ll go as follows:
The elite – These are the true cream of the crop of the NFL, the best of the best. Seven fit here.
The solid – These are guys I would define as good players that you are happy to leave in as your team’s QB for five-plus years. They may not be the best QBs out there, but they’re good enough to give your team a chance. I have 13 in this category.
The young guns – These are guys playing in their third or fourth year and are on a trajectory to fall somewhere in either the elite or the solid groups in the next few years. There are six of them.
The bad – These will range between flat busts and QBs that are starters for their teams, but are unlikely to hold that job long because they’re not good enough to rely on being winners in the long term. There are 14 in this category.

There may be some discussion as to whether certain guys could fall into the “bad” vs. the “solid” category. And I’m fine with that, but in general, I don’t see a lot of argument there. We may also argue between elite and solid a bit to, but I think that’s an even less important distinction, as the Rule isn’t really called out to distinguish between the good and the great; I’m just looking at it for general instruction.

And so the meat of the discussion, how does the Rule really perform?

Elite, fits Rule
Brees, Manning, Rivers
Elite, doesn’t fit Rule
Brady, Favre, Rodgers, Roethlisberger
Solid, fits Rule
Eli Manning, Kolb, Schaub, Fitzpatrick, Romo
Solid, doesn’t fit Rule
Palmer, McNabb, Cutler, Kerry Collins, Cassel, Hasselbeck, Vick
Young gun, fits Rule
Matt Ryan
Young gun, doesn’t fit Rule
Henne, Flacco, Freeman, Sanchez, Stafford
Bad, fits Rule (as in, does not have 26 and 27 and 60)
Alex Smith, Quinn, Leftwich, Carr, Losman, Russell, Campbell, Harrington, Boller, Ramsey, Tarvaris Jackson, Vince Young
Bad, doesn’t fit Rule (as in, does have 26-27-60)
Leinart, Grossman, Orton

One thing worth noting here, I count Aaron Rodgers in the “doesn’t fit Rule” category, because I don’t count his starts at community college as true college starts. If you do count those, then he fits the Rule. However, starting at community college seems to me to be a terrible measure, as I don’t know that the NFL has ever seen a community-college-only QB ever come into the NFL and succeed. I’m not sure one has ever entered the NFL, and either way it doesn’t impact the overall numbers much.

And so, what percent of each group has the Rule correctly predicted?
Elite – 3/7 = 43%
Solid – 5/12 = 42%
Young Gun – 1/6 = 17%
Bad – 12/15 = 80%
Overall – 21/40 = 53%

Let’s look at it a different way. What percentage did achieve 26-27-60 and succeed and what percentage did not achieve 26-27-60 and turned out bad?

Did achieve 26-27-60 and succeeded – 9/12 = 75%
Did not achieve 26-27-60 and were bad – 12/28 = 43%

It looks like the Rule is actually pretty solid at predicting that a quarterback will succeed on some level if they fit the criteria. However, the Rule appears to be a terrible predictor of failure if the prospect doesn’t fit all the criteria. It also appears to be poor at distinguishing exactly how good a QB is going to be, as only 33% of those that achieved 26-27-60 turned out to be in the elite or young gun category, while 32% that didn’t achieve 26-27-60 fell into elite or young gun.

Overall the Rule doesn’t do a much better job at predicting overall success or failure of the population pool than simply flipping a coin. I hated this Rule from the first time I read about it, because Lopez cherry-picked his quarterbacks to fit his study and never really looked at how good of an indicator it was across the whole population. Seeing the numbers in greater detail doesn’t change my opinion.

1 comment:

  1. Interesting. Although I didn't have all the data, other than what was available on the SI site, I looked at "outliers" in the three data sets based upon standard deviations. When a QB was a standard deviation above the mean on at least one area AND not below in any area, they tended to be successful (in fact none could be described as unsuccessful-- at least those from the SI story). Of the elite, only Brees wasn't a standard deviation above the mean in any one area-- but he also wasn't below in any of the areas When a QB had at least one Standard Deviation below the mean (and none above the mean) they tended to be failures. Leaf, Harrington, Smith, Young, Vick and Flacco all this this profile. It is fair to say that Vick and Flacco are far from busts, but no statistical methodology will account for all the variables (such as perhaps Flacco's size and surrounding teammates and Vick's mobility). Thus, perhaps, if the data is sliced a different way, it may have even more utility.

    ReplyDelete

About This Blog

Twitter: oblong_spheroid

  © Blogger templates The Professional Template by Ourblogtemplates.com 2008

Back to TOP