STDL: Broncos upgraded a bit at QB in 2012

I wanted Tim Tebow to be a great NFL quarterback.

No, really. I did.

As Tebow's Hail Mary pass to Brandon Lloyd fell incomplete, thus ending Denver's wretched 2010 season, I was the one Broncos fan at SideBar in New York City (the birthplace of Tebowing) openly cheering the result.

Now, you're not supposed to root against your team, but Denver's draft standing was at stake - a win would have dropped the Broncos all the way from second to fifth in the 2011 Draft, and as much as we'd like to think that pick may have turned into Aldon Smith or J.J. Watt, we can be sure it would not have netted Von Miller.

Draft position aside, there was nothing to be gained out of finishing 5-11 rather than 4-12, and besides, I thought something much more important had occurred that day at the Big IF.

As I prematurely proclaimed the next morning - two years ago today - I thought the Broncos had found a quarterback, one who would be leading the team "for the next decade or so." I even took the extra measure of comparing Tim's touchdown rates and YPA with those of Tom Brady and Philip Rivers.

Whoops. Tiny sample alert.

Doubts started to creep in for me during the following preseason, and I thought the decision to go with Kyle Orton as the starting quarterback was not an overly difficult one. I was also more than fine with the decision to give Tebow a shot after Orton had looked like a shell of his former self over the first four and a half games of the season.

Of course, the Broncos, and Tebow, started winning a couple weeks later, and Tebowmania had arrived in Denver for real.

But the throws kept hitting the turf, or sailing out of bounds, and frustrated receivers would run fruitless routes, getting open only to watch Tebow scramble for a minimal gain or throw the ball in the general vicinity of a different, less open teammate.

Tim wasn't throwing the ball very often, but the more he did, the more apparent it became to me that the guy just couldn't throw a football with consistency for the life of him.

All along, I'd been working on a quarterback stat that would take running the ball into consideration. After all, with Tebow and Cam Newton in 2011, and RG3, Russell Wilson, and Colin Kaepernick this year, we've been seeing versatile passers who also threaten defenses with their legs - moreso than even the great John Elway had done back in the day.

But with Tim having thrown such a small amount of passes, I didn't want to really start discussing his numbers until the sample had grown. I'm no statistician, but I realize that the worst thing you can ever do with a statistic is to overstate its significance. Writing that Tebow was on his way to becoming a great NFL quarterback just felt premature. Yes, more early than declaring him Denver's QB of the future had been the prior January.

Then, with Denver's win streak at four games, during which Tim had completed just 30 passes for a 44.8% completion rate, statistical hack Kerry Byrne dropped a bomb over at Sports Illustrated. Through his unconscionable abuse of the basic rules of mathematics, Byrne had determined that Tebow was protecting the football better than the league's best quarterbacks, and was producing touchdowns at a faster rate than them as well.

Indeed, Tim had been faring well at both scoring touchdowns and not turning the ball over. But he wasn't getting the ball down the field, and therein lied the problem. Touchdown and turnover rates fluctuate from year to year and depend largely upon luck, while things like yards per attempt and sack rates do not. Without exception, all of the great quarterbacks fare well relative to their peers in YPA and the derivative NY/A (Net Yards per Attempt, which includes sacks and is the most predictive stat of future success) and ANY/A (Adjusted Net Yards per Attempt, which bakes TDs and INTs into YPA to provide a retrodictive measure of what's already happened) figures from PFR which we often cite here.

Byrne's statistical sacrilege prompted me not only to respond to his column with one of my one, but also to unveil a few days later the QB stats that I'd been working on.

I know, stats are for losers, and all Timmy did was win.

But unfortunately, the more the tape piled up, and the statistics went from tiny samples, to small, to medium ones, it became more and more clear that Tebow had no chance of becoming a great NFL quarterback. By the end of the regular season, he ranked 28th among quarterbacks in Adjusted Net Yards per Touch, and 36th out of 40 in Net Yards per Touch, which like its PFR counterpart NY/A, is meant to be a predictive measure.

364 days after my last post about QB stats, it feels like the right time to examine some from the 2012 season.

After all, the regular season has concluded (compiling the stats on a weekly basis is a mountainous and prohibitive chore), and our old friend Byrne is back with another stupid column saying how great Tebow is, using his site's mathematically-challenged Real Quarterback Rating* as backing. Byrne's assertion is that the seven coaches and five general managers who were fired this week would have all kept their jobs if there were nine Tim Tebows to go around.

*The problem with traditional QB rating is that it overcredits completion rates. CHFF's version compounds that by treating rushing attempts like completions, thus turning a flawed metric into a completely useless one.

Let's all laugh at that one together, and then remember that the Broncos have Peyton Manning.

Now, we all know that Peyton is a great quarterback. We don't need no stinking stats to figure that out, do we? But he doesn't run the ball, and many of the other top passers currently in the league do so - namely, Aaron Rodgers, Newton, RG3, Wilson, and Kaepernick.

How does that added element affect things? Are any of these players more lethal than Manning due to their running ability?

Simply, no. Not in 2012, at least.

Let's take a look at the figures. As always, the math:

We're basing our formula on the work already done by Chase Stuart for the PFR Blog (now of the fine footballperspective.com) but adding some tweaks. Here's the original formula:

[(PYD + 20*(PTD + RTD) - 45*INT - SKYDLST - 35*(FUM-FumRec)) / (ATT + SK +RTD)]

The values being used to bake in Touchdowns, Interceptions and Fumbles were based upon the work done by the authors of The Hidden Game of Football, which had originally valued a Touchdown as being worth 10 yards. Stuart explains the increase to 20 yards here. Here's the formula we're going to use for our Adjusted Net Yards per Touch:

[(Pass Yds + Rush Yds + Rec Yds - Sk Yds + 20*(Pass TD + Rush TD + Rec TD) - 45*INT - 35*FUM) / (Pass Att + Rush Att + Rec + Sks)] 

The differences between our formula and that used by Stuart at PFR are as follows:

1. We've added rushing and receiving yards, and receiving touchdowns. As for the receiving stats, hey - you never know right?
2. We're going with straight fumbles rather than net fumbles, because we're treating all fumbles as negative events - whether or not a team recovers its own or not. As we all should know by now, fumble recoveries are random events.

The Negative Play % column shows the frequency that each quarterback is sacked, intercepted, or fumbles - in other words, the worst results of all.

First, let's look at ANY/T, which, as mentioned above, will be one of those metrics (like ANY/A) better at telling a retrodictive story than predicting the future (because pushing the ball down the field is a more easily repeatable skill, while touchdowns, fumbles and interceptions are more random):

2012 QB Data through Week 17, sorted by Adjusted Net Yards per Touch

Manning led the league in ANY/T this year, and it wasn't even close. Tom Brady, another traditional/non-running QB is second, and nobody comes close to him there. Aaron Rodgers is a distant third, almost entirely because he took more sacks than Manning and Brady combined.

Of course, Denver's season isn't yet over, so more important than looking at ANY/T and measuring what Peyton has already done for Denver, let's examine how he stacks up in terms of Net Yards per Touch, which is as follows:

[(Pass Yds + Rush Yds + Rec Yds - Sk Yds) / (Pass Att + Rush Att + Rec + Sks)]

This is simply a matter of adding rushing and receiving stats to NY/A. Here are the results:

2012 QB Data through Week 17, sorted by Net Yards per Touch

No, that wasn't the same chart. Yes, Peyton also led the league in NY/T, although by a smaller margin, and with Kaepernick posting the second best rate.

It's rather simple. We don't need to even delve into the backstory of Manning uprooting his life and joining a new team, with mostly unfamiliar players, and overcoming what had been considered a career-threatening injury. We'll save that all for the MVP arguments.

What the stats say is that Peyton Manning was the very best quarterback in the NFL in 2012.

A few more observations:

1. Manning (5.42%) ranked second in the league in Negative Play Rate, just a shade behind Brady (5.39%). However, it's only the sixth best such figure of Peyton's career.
    
2. All three quarterbacks Denver has a chance of facing in the divisional round rank 19th or worse in NY/T.
    
3. Including Peyton, five of the top six QBs by NY/T are in the playoffs, but Denver won't see any of them prior to the conference championship (Brady).
    
4. Eight of the league's top thirteen passers by NY/T are leading playoff teams, and adding Alex Smith would make those figures read nine of fourteen.
    
5. Tebow's 2011 stats would have placed him 32nd in ANY/T, and 35th out of 36 quarterbacks in NY/T.
    
6. Tim's wretched numbers were often excused away by his backers as owing to his youth and inexperience. While that may have been a valid argument, the presence of Griffin, Kaepernick, and Wilson among the top ten in both ANY/T and NY/T this season would stand as proof that the NFL learning curve isn't quite that steep.
    
7. Thank You, John Elway.