The New York Times recently published award-season predictions by Chase Stuart, who “writes about the historical and statistical side of football at his site, FootballPerspective.com.” The post includes his pick for NFL Offensive Rookie of the Year, Robert Griffin the third. Obviously, I’d like to see Russell Wilson get it. That said, I understand that RWI probably isn’t going to win in a season when there is a strong case to be made for Andrew Luck, RGIII, and even Alfred Morris.
But here’s the thing: if there is a case to be made for RGIII, Stuart didn’t make it.
I plan to quote the entire (short) section more-or-less in full, so no need to check out the original:
One of the most difficult awards to pick. Griffin, Andrew Luck and Russell Wilson had excellent seasons.
So far, so good.
Next, Stuart turns to Andrew Luck. To my mind, the “Colts are most-improved thanks to their QB” sounds like an argument for MVP, but I’m just here police internal consistency, not weigh the merit of Stuart’s arguments:
Luck had the least support and was asked to do the most, as the Colts placed the full weight on his shoulders. He was not able to rely on the zone-read or the threat of play-action, and was constantly making difficult throws downfield. While that hurt his completion percentage, his ability transformed the Colts from a 2-14 team to an 11-5 playoff squad, and Luck led the N.F.L. with seven game-winning drives in the fourth quarter or overtime.
Next man up is Russell Wilson, and Stuart makes his case on the strength of traditional football counting stats:
Russell Wilson tied Peyton Manning’s rookie record with 26 passing touchdowns and chipped in with four scores on the ground. He ended the season with a 100.0 passer rating, breaking Ben Roethlisberger’s rookie record. He helped turn the Seahawks into maybe the most dangerous team in the N.F.C., and he improved in the second half of the year. In most seasons, he’d be an obvious candidate, but this year, he’s forced to play third fiddle to Luck and Griffin.
Finally, we come to the Griffin pick, which I’ll break it into two parts. First, why RGIII deserves the award over Wilson:
So why will Griffin win the award? Here’s the key statistic: RGIII finished the season first in the N.F.L. in yards per pass attempt and first in yards per rush attempt. And while Wilson may have broken Roethlisberger’s record, Griffin broke Wilson’s record a few hours later, finishing the season with the highest passer rating by a rookie in N.F.L. history.
Still — so far, so good. Where Stuart originally laid out Wilson’s credentials in terms of counting stats, he follows up by pointing out that RGIII’s were better. Accepting the importance of counting stats to the awards process, I have no problem with this argument. It may not be right, and I may not agree with it, but it’s at least internally consistent.
Now we come to Griffin over Luck, and this section also more or less works out:
Washington finished fourth in points scored, and Griffin guided them to the playoffs against a much tougher schedule than what Luck navigated. He has significant edges over Luck in most key passing categories, and he led the N.F.C. in Adjusted Net Yards per Attempt.
Think back, for a moment — or just scroll up — to Stuart’s argument for Luck. Recall that it focused on context in the abstract. Stuart pointed out that Luck carried the Colts on his back all season long. In knocking him back down, Stuart again turns to context, but this time cites more specific statistics: RGIII played against better defenses, and had more adjusted net yards, the product of (gasp) a mathematical formula. As in the Wilson-Griffin debate, the arguments line up: Stuart limits the debate between Griffin and Luck to individual context.
But when you take a step back and look at the bigger picture — yes, in context — it’s difficult to avoid the conclusion that Stuart’s process is fundamentally flawed. Specifically, Stuart dismisses Wilson on the basis of counting stats, and Luck on the basis of context, but never considers either player on the terms by which he measured the other.
Consider Luck on his raw statistics. He racked up over 4,000 passing yards and 23 touchdowns, but also threw 17 interceptions. RGIII threw just under 3000 passing yards, 20 TDs, and only 5 interceptions. Applying the “Wilson standard” to Luck, you can still make the case that Griffin was better — especially when you consider his rushing totals — but Stuart never makes that case.
But now consider Wilson in context — specifically, using the “tougher schedule” criteria Stuart used to knock out Lock and award the trophy to RGIII. Funny thing: when you adjust the QB stats by strength of schedule, Wilson was the better player by a significant margin. In fact, according to Football Outsiders, Russell Wilson just put up the best statistical season for a rookie since 1991:
Wilson surpasses Matt Ryan for the top rookie quarterback season in FO’s database, and becomes the first freshman at the position to surpass the 1,000-DYAR* barrier.
Here’s how the rankings shook out:
1) Russell Wilson, Seattle Seahawks: 1,014 DYAR (867 passing DYAR, 147 rushing)
2) Robert Griffin, Washington Redskins: 838 DYAR (729 passing DYAR, 109 rushing)
3) Andrew Luck, Indianapolis Colts: 379 DYAR (255 passing DYAR, 124 rushing)
4) Ryan Tannehill, Miami Dolphins: 39 DYAR (37 passing DYAR, 2 rushing)
5) Brandon Weeden, Cleveland Browns: -266 DYAR (-290 passing DYAR, 24 rushing)
In the context of the league as a whole, Wilson ranked 8th; RGIII was 12th. This shouldn’t be all that surprising, given the specific teams Wilson and RGIII were asked to face. Here are the four best pass and run defenses the Redskins took on, followed by their ranking (according to Football Outsiders), followed by the comparable opponents of Seattle:
Washington’s pass defenses faced:
St Louis (8th)
Washington’s run defenses faced:
Tampa Bay (3rd)
St Louis (10)
Seattle’s pass defenses faced:
Arizona (2) x2
San Francisco (6) x2
Green Bay (7)
Seattle’s run defenses faced:
San Francisco (2) x2
New England (6)
With all due respect to RGIII, Wilson had a quantifiably rougher introduction to the league. Or if you prefer Football Insiders, here’s an extended quote from ESPN:
The Total QBR metric we’ve consulted in evaluating quarterback play can be tweaked to account for strength of opposing defense.
Alok Pattani of ESPN’s analytics team passed along information showing how these adjustments would affect QBR rankings for the 36 quarterbacks with enough plays to qualify for consideration.
Seattle’s Russell Wilson and St. Louis’ Sam Bradford joined Detroit’s Matthew Stafford as the biggest winners in terms of ranking spots gained. Each would move up three ranking spots if opponent strength were factored.
Wilson jumps from eighth to fifth, moving past wild-card playoff opponent Robert Griffin III. Griffin moved back one spot to No. 7.
The point isn’t necessarily that Wilson should be Offensive ROTY, or that you have to buy the specific statistics cited; the point is neatly summed up in what comes next from ESPN:
The numbers suggest Griffin and Wilson are interchangeable from a production standpoint.
And that’s what happens when “the numbers” are context-dependent. If context (specifically, opponent defenses) is enough to push Griffin over Luck, it should also be enough to push Wilson back into the conversation.
It’s certainly convenient to cherry-pick the class of statistics you consider while trying to make the case for a particular player, but it’s hardly objective. Granted, this is sports, it’s not supposed to be objective, and it’s not worth getting upset over. But it’s also sloppy reasoning published in the paper of record, and that is worth getting upset over. Journalists are entitled to their own opinions, but I don’t think it’s too much to ask for their analysis be at least internally consistent.
It’s here that I will obligatorily note that the Times is an East Coast media institution, but then again ESPN pointed out that RGIII and RWI are pretty much the same player, so that wouldn’t be entirely fair.
It should be a good game this Sunday. Go Hawks.
*DYAR stands for Defense-adjusted Yards Above Replacement — the key term here being “defense-adjusted yards”. The “above replacement” half simply references the stat’s baseline; you can read more about it here.