*A new system of rating QBs puts Aaron Rodgers on top and leaves RG3 some room to grow.*

How can we know who the best NFL quarterback is at any given time? There are plenty of ranking systems trying to figure that out, and on TheAtlantic.com recently Allen Barra proposed another one: the usage of adjusted yards per pass attempt (AYP) as the best available metric to define quarterback performance. Allen makes the point that existing ranking systems are "irrelevant" and that statistics such as the NFL's quarterback rating and ESPN's Total QBR are needlessly complex for the average fan.

While he has a point that these kinds of ranking systems have their flaws (doubly so for Total QBR), his proposal fails to capture a more important factor when trying to measure the sum of a quarterback's value to his team. AYP combines two football statistics: yards per attempt and interception percentage. It's true that Adjusted Yards per Pass could be a useful metric because both input statistics can indeed be key indicators of a quarterback's skill level. But there is nothing in Barra's formula that points to a quarterback's ability to score a touchdown. The mission to come up with a formula that does leads to a different key metric: the difference between touchdowns and turnovers committed.

Above all other things, it is the responsibility of the quarterback to maximize the number of touchdowns scored and minimize the number of turnovers committed. The TD-INT ratio is a widely used metric, but interceptions don't tell the whole story. Football analysts need to stop ignoring the effect of quarterback fumbles on the effectiveness of an offense. An interception most of the time places the opposing team down the field, past the original line of scrimmage. Fumbles on the other hand tend to happen in the pocket or near the line of scrimmage, making them potentially far more damaging to the offense.

I propose the following formula:

Total fumbles are divided in half to account for the fact that, over a stretch of time, a team will recover roughly half of their fumbles. In addition, I've added the number of fumbles to the denominator to account for the fact that a fumble is not counted as a pass attempt like a touchdown or interception is. Although it is less useful on a single-game basis, the touchdown-turnover percentage (TD-TO%) provides a more complete analysis of a quarterback's value to his team over a longer stretch of time. In particular, it casts a very different view on Robert Griffin III (leading in AYP at the time the article was written), the player Barra highlights as the prime example for why AYP works.

I had the opportunity to see him live at the Redskins game against the Carolina Panthers. It was a rough game for the Redskins (they lost 21-13), and although there were many comments from fans during and after the game as to why their team has underperformed, one in particular stood out for me. About halfway through the third quarter, a young fan sitting near me asked his father, "When is the last time the Redskins scored a touchdown? They can't get the ball into the endzone!"

The mission to come up with a formula that truly points to a quarterback's ability to score a touchdown involves a different key metric: the difference between touchdowns and turnovers committed.

The key metric here, important for measuring Griffin's progress as a rookie, is that he is one of only four quarterbacks so far this year to average less than one passing touchdown per start: The others are Ryan Tannehill, John Skelton, and Cam Newton. Furthermore, Griffin has had an alarming tendency to put the ball on the ground this year, fumbling eight times (two lost) in nine games. Using my formula above, RG3's true TD-Turnover ratio is 8-7: His TD-TO% of 0.37% (Ranked 19th overall) is closer to that of Christian Ponder (0.17%) than Aaron Rodgers (5.95%).

This metric shows Griffin to be far from the upper echelon of the NFL, but I think most would say that's a fair assessment. Griffin is a talented rookie with a lot of room to grow and (unfortunately) a sub-par supporting cast. With an improvement of the Redskins team and a couple years of experience he could certainly be one of the top five quarterbacks in the league. However, as it stands now, Barra's metric fails the test provided by Robert Griffin III.

As with any ranking system, the proof should be in the results. According to the TD-TO%, the top five quarterbacks through week nine are Aaron Rodgers, Peyton Manning, Tom Brady, Matt Ryan, and Drew Brees. In contrast, Barra's AYP produces the following top five: Peyton Manning, Josh Freeman, Tom Brady, Robert Griffin III, and Matt Schaub. Personally, I'd put Matt Ryan over Josh Freeman any day of the week, although Barra might disagree.

Is there room for debate? Of course there is. I don't put much stock in analysis that claims to be the be-all, end-all in any field. Thankfully the evolution of statistics in football has allowed the proliferation of many metrics (including TD-TO% and AYP) that should be combined to tell a broader story of a quarterback's performance. One could even use championships, as Barra says, to determine a list of the best quarterbacks of all time (though I'm sure that Dan Marino would take offense to being placed behind the immortal Jim McMahon). Ultimately, since there is no "better" stat, I invite you to come up with one of your own.

This article available online at:

http://www.theatlantic.com/entertainment/archive/2012/11/and-now-another-effective-way-to-rank-nfl-quarterbacks/264978/