Increasingly pass-happy NFL might not be


Perhaps you've fallen for the line about NFL teams becoming increasingly pass-happy on offense. Turns out that is not necessarily the case.

It's a little embarrassing, in retrospect, to think how many times this notion has appeared as fact in these parts recently.

There was my reference from the NFL scouting combine, and another two months later in relation to how teams are valuing players at certain positions. Another mention hit the blog last week. Then came the reference that led Mark from San Francisco to hit the NFC West mailbag with my favorite kind of feedback, the smart kind.

Highest Dropback Rates Since 1990

Several exchanges over email led to a different conclusion: Teams are passing more successfully, but not necessarily more frequently.

"A sports blog isn't exactly the place for this type of lesson," Mark wrote, "but as fans look more and more at stats, it can be easy to see meaning when the differences are simply 'in the noise.' "

This all began with the item posted here Sunday in response to a request for information demonstrating the "inflation" of passing numbers over the past decade or so. I put together a table showing passing yards per game, yards per pass attempt, pass attempts per game, and NFL passer rating for each season since 2000.

Most of the higher numbers were posted since 2007. This seemed to affirm the widely accepted premise of increasing pass-happiness.

But as Mark explained, most of the averages since 2000 did not deviate enough for the differences to be statistically significant. That was especially true once we determined, at Mark's request, whether pass dropbacks were increasing as a percentage of total plays, and in a manner statisticians would find significant. They were not.

We expanded the time frame to 1990 in an attempt to get a bigger-picture view. Using this new timeline, we determined passing yards per game have increased significantly (beyond one standard deviation for you math heads) in each of the past three seasons.

But teams continue to drop back for passes between 54 and 58 percent of the time, same as they have done for two decades. The rate was 57.1 percent of plays last season, same as it was in 1999 and less than the rate in 1995. The first chart shows the 10 highest dropback percentages since 1990 (pass attempts plus sacks, divided by plays). Note: I removed a mention to 1991 after realizing the dropback percentage for that year was lower than initially thought.

Teams appear to be throwing slightly more frequently over the past few seasons, but we'll need a few more years of increases to know the meaning. The statistically significant gains have been made in other areas, notably completion percentage and NFL passer rating.

Highest NFL Passer Ratings Since 1990

"What HAS changed in the past few years is the SUCCESS of passing -- four of the past five years are statistically above the average since 1990 by plus-one standard deviation in terms of completion rate," Mark explained. "This almost certainly translates to the increase in passing yards over the past three years. Also note that NFL passer rating has risen significantly over the past five years and is probably linked with the increased yardage."

Rules changes have also enabled superior passing.

The increase in quality could reflect, to some degree, a stronger group of quarterbacks recently. ESPN's John Clayton sees value in this reasoning. He points to 2004 as a pivotal year, because that was when several top passers either entered the NFL or began to hit stride. Ben Roethlisberger, Philip Rivers, Eli Manning, Matt Schaub, Drew Brees, Tony Romo and Carson Palmer come to mind. Peyton Manning and Tom Brady were already established, but also rising.

If that were the case, we might expect to see the quality of passing stats diminish as these players retire. But we might not see any change in how frequently teams drop back to pass.

Dropback rates spiked in the mid-1960s, fell in the 1970s and then jumped markedly beginning in the late 1970s and continuing into the early 1990s, according to research conducted by Alok Pattani of ESPN's analytics team.

"Dropback percentage has increased by about 0.06 percentage points per year from 1990 to 2011," Pattani said. "The slope of this regression is borderline statistically significant. However, if you just look at 2000 to 2011 or even 1995 to 2011, the increase in pass play percentage over those years is not statistically significant.

"So, while dropback percentage does appear to be on the rise, the increase doesn’t appear to be significant (i.e., it could be random) relative to the last 10 or 15 years."

Of course, there are other ways to measure pass-happiness. Are teams passing more frequently on first down? In first halves? When leading? Are they passing more frequently in some traditional running situations? Are they favoring personnel groupings with more wide receivers?

These are questions for another time. For now, Mark has provided a valuable reminder. Teams might be increasingly happy about their results when passing, but that doesn't make them increasingly pass-happy.