The Pro Bowl has become a yearly debate over whether All Star games matter in any tangible way. Fans throw out suggestions every year for how the game could be improved, but mostly the NFL just changes location. Using the Pro Bowl as a measure of success, due to the silliness of the game itself and the sometimes dubious nature of selections to it, is sometimes scoffed at. I, for one, think it is an accurate measure of player success since, while there may be some players that get snubbed every season and a few who make it in on name alone, the vast majority of Pro Bowl players are very good at football. Due to availability of data, it was the first "success metric" that I used when I began comparing measurements and NFL impact.
The effectiveness of any metric is often gauged by where it doesn't apply. If, for instance, you were to look at something as simple as the 40 time as a measure of success, the first player mentioned would be Jerry Rice and his 4.7 time (despite not being a real thing). So when I first created RAS, the Relative Athletic Score, the initial push-back was for the players who scored low but still found success. What I tend to use metrics for, and I hope everyone would, is to chart trends. Probability, not possibility. There are so many factors involved, there will never be a 100 percent accurate metric, no matter what some popular analytics sites may say.
This is probably a good time to mention, as I often do, that RAS should never be used to point directly to success. A high score does not mean success in that regard (which would be causality) just like being a tremendous athlete doesn't immediately mean superstar NFL player. Instead, we look at who has been successful, then compare where those players won and find (generally) that those who had success also had high scores. That's correlation. We look in metrics to find correlation, not causality. Causality would be great, but like rational Bears fans I'm not sure it really exists.
It should come as no surprise that quarterback has no clear correlation with RAS and success for Pro Bowls. Tom Brady's putrid 0.46 out of 10 score ranked 251st out of 262 QBs (if he weren't tall, it'd be last). Roughly half of the Pro Bowl QBs since 2000 have ranked above 5.00 RAS, with the rest below. You have some top tier athletes who have made Pro Bowls, like Cam Newton (10.00 RAS), Robert Griffin III (9.85), Andrew Luck (9.66), and Russell Wilson (8.44), but you'll also note that they did a ton of damage with their feet as well as their arms. Meanwhile, on the other end, you have guys like Brady, Andy Dalton (2.48), Tony Romo (3.21), and Matthew Stafford (3.70) who also made Pro Bowls, but weren't super athletic. All in all, RAS has almost the exact same correlation to success (using Pro Bowl) as the 40, vertical, or literally any individual metric does for QB. Which is none.
More than twice as many RBs who made a Pro Bowl had a 5.00 RAS or higher than those who had a score below average. Out of the 33 RBs who both qualified for RAS and made a Pro Bowl, a whopping 13 of them had a score of 8.00 or higher. That's almost as many as had scores from 0.0 to 6.0 (14). It shouldn't be surprising that guys like LaDainian Tomlinson (9.46), Adrian Peterson (9.12), and Chris Johnson (9.56) had incredible scores, but there were some surprises at RB. Marshawn Lynch, for instance, was sandwiched between Darren Sproles and Maurice Jones-Drew with only a 5.09 RAS. It was also interesting to see guys like Frank Gore (3.40), LeSean McCoy (4.94), and Devonta Freeman (0.76) below that average line. RAS has a higher correlation than any individual measurement for RB, though only slightly when compared to 40 times.
No position has shown a higher correlation than tight end. Every single tight end drafted in the first round since 2000 has had a RAS higher than 5.00, while only two of the 18 Pro Bowl TEs over that span of time rated below average. Four-time Pro Bowler Alge Crumpler scored a 4.18 while Zach Miller and his 3.77 score made a single Pro Bowl. It should surprise literally no one that the likes of Jordan Cameron (9.97), Vernon Davis (9.94), Jimmy Graham (9.65), and Travis Kelce (9.03) all rated in the elite. In fact, half of the 18 Pro Bowl TEs scored an 8.50 or higher. And yes, for those wondering, Eric Ebron scored well above average (6.86).
I briefly hit upon this topic when I looked at the offensive lines for the entire NFL, but offensive tackle is another position that has a huge correlation with RAS and success. My previous numbers were slightly off, as there were a couple players who were listed at guard who actually made the Pro Bowl as a tackle, but we're still looking at 21 out of the 25 qualifying OT who scored 5.00 or greater. All four of the Pro Bowl OTs who scored below average made only a single Pro Bowl. Those are Marvel Smith (3.70), Michael Roos (3.52), Tyson Clabo (2.12) and David Diehl (1.22). All four of those players were drafted in 2005 or earlier. 15 Pro Bowl OTs rated an 8.00 or higher for RAS, or 60 percent. One thing that surprised me is that the four OTs who scored below 5.00 were the only players who had a 40 time below average. That's right, RAS has the same correlation in this instance as 40 time does, which apparently matters more than any other individual metric. Who would have guessed that? Despite so many players above 8.00, only three scored above 9.00. Joe Staley (9.87), Trent Williams (9.35), and Tyron Smith (9.22) headline that elite group.
Guard is another position I touched on briefly before, but since I had to shuffle a few players the numbers are slightly different. 15 out of the 19 offensive guards who made a Pro Bowl since 2000 had a RAS higher than 5.00, and 10 of those had a score higher than 8.00. Evan Mathis (9.97), Steve Hutchinson (9.75), and Kyle Long (9.53) lead this group, while only Leonard Davis (4.60), Mike Iupati (4.41), Jahri Evans (3.79), and Shawn Andrews (1.55) had a score below average.
Center is the only position on the line that has no clear correlation with RAS to NFL success. That's a little odd, but my guess is that many of the talents for center are not measurable under any current metrics. While over half of the centers who have made a Pro Bowl since 2000 were above average RAS, it is not enough to make a clear correlation. What was interesting to me, however, is that of the six centers that rated below average, only one (Dan Koppen, 3.82) was even close to an average-sized center. The other five were all huge. You have the Pouncey brothers (Mike 3.15, Maurkice 2.79), Max Unger (3.39), and a pair of Cowboys centers in Andre Gurode (3.33) and Travis Frederick (1.27). We only had two guys in the elite, Jason Kelce (9.39) and Chris Myers (9.52). Taylor Decker and Larry Warford's mentor, LeCharles Bentley, notched a nice 5.76 for his RAS.
Everyone's favorite group to pick the outliers from. Wide receiver has some very notable exceptions that people love to point to and explain how pointless measurements are. In fact, players with a RAS over 5.00 make the Pro Bowl just under twice as often as those with scores below average. Still, it has the third-lowest correlation for offense, behind QB and OC. Despite that, the same number of players scored an 8.00 and above as scored below 5.00, 13 out of 37 receivers. This is a good time to point out that our highest scoring Pro Bowler, Calvin Johnson, also has the most impressive RAS out of the 5,400 players and counting in my database. Calvin's raw average (a behind-the-scenes number, just an average of all of the individual scores) was 9.68, almost a full point ahead of the next highest Pro Bowler, Andre Johnson, though you'd never know it looking at just their RAS where Calvin is 10.00 and Andre is 9.94. Players like Julio Jones (9.86) and Roddy White (9.07) played across from one another while Emmanuel Sanders (9.00) and Mike Wallace (8.79) were once a thing. Speaking of Steelers receivers, Antonio Brown is one of our great outliers at only 1.28 out of 10. Joining him near the bottom are, in order, Randall Cobb (0.82), Brandon Lloyd (0.49), Wes Welker (0.47), Anquan Boldin (0.42), and Jarvis Landry at only 0.04 out of 10.
RAS has shown to have a pretty good correlation for offensive players as a whole and NFL success, using the Pro Bowl as a measure. You can check out the entire list of Pro Bowlers and their RAS here. If you have any questions or complaints about RAS, let us know in the comments or you can always ask me directly on Twitter @MathBomb.