Weirdly Depressed Offensive Years in Pro Football since the USFL Era Passing Explosion

Last week, I looked at NYPA progression over the post-merger era specifically as an indicator of the overall passing strength of a given era.  The interesting conclusion is that NYPA draws a completely different picture of the fertility of passing eras, lumping 1978-2006 as one era where NYPA = 5.8, up from the mid to lower fives of the depressed seventies, but completely missing the wide open-ness of passing in the eighties because sacks were up and interceptions were not down.  Clearly, a number of different measures can show us how different passing was in the mid eighties, even if the value metrics of record show that league-wide passing efficiency hadn’t gone anywhere.  Yards were just in greater supply, for a limited time, presumably until offensive coordinators realized that they were also inflating the opponents’ point totals as well as their own.

Well, my research also showed that while the average points per NFL game has generally increased since the madness of the 1980’s, there have been three one-or-multiple year dips in scoring over the last 25 years, at least dips not caused by factors such as lost games or replacement players.  Football scores and stats have been remarkably consistent in their general, “slightly upward” trend, which just makes me want to mess around and examine the exceptions to the rule.  This is my chance to do just that.

1991-1993

The average NFL quarterback in the late 80’s was a classic gunslinger, downfield throwing, play action based passer or maybe a run and shoot, high efficiency, high risk, throw to win kind of guy.  That was great and all, and average in any era had some sort of value to a franchise, but if the passing game was ever truly going to replace the running game as a viable means of moving the sticks, turnover rates could not be sustained as high as they were in the 1980’s.  To date, throwing the ball had always risked turnover rates in the 1-in-17 ish range based on sack-strip fumbles and interceptions.  But the forward pass got by for a long time as a high-risk, high-reward play because the status quo was the run the ball thirty or more times a game.  In the 80’s, rushing rates dropped and passing rates increased, which meant that teams were turning the ball over at really incredible rates.

Because scoring totals were high for both teams, this actually went on for years until coaches started to realize that offenses weren’t actually getting any more efficient.  They were making defenses more productive in the process by simply forcing a lot more high variance outcome via the pass.  Now, if you were the Houston Oilers or Washington Redskins, great, you weren’t an average team and defenses weren’t going to stop you an average amount of time, and the yards were there for your receivers, and you might as well use them.  But for the average team, gross efficiency wasn’t any different in 1988, 89, or 90 than it was in 1978, 79, or 80.  The secret for increased was thought to be superior offensive thinking and QB play when it was really just shorter fields and the slightly inflated value of a yard.

These three seasons saw a correction in that kind of thinking.  Scoring dipped below 19 PPG to about 18.7 for the first time since 1980.  Average yards per attempt dipped below 7 for the first time since about 82.  Adjusted net yards per attempt though actually increased.  Defenses weren’t more prepared to stop passing attacks in the early 90’s, despite the more cautious approach by offenses.  If defenses were better suited to stop passers, there wouldn’t have been a ANYPA increase.  But as the value of a sack and interception started to take root, players who had high interception and sack totals started to lose jobs to veterans who could get the ball out to the right team.

But completion percentage and yards were down, and since the conservative thinking of these years didn’t exactly result in coaches adopting an aggressive scoring strategy, scoring fell.  Beyond that, the “heroes” of the late 80’s started to lose their footing in the 90s.  Bernie Kosar was gravely affected by the shift.  Joe Montana took more negative plays (defined as sacks and INTs) in the final three years as a starter for the 49ers than he ever did during the height of the Bill Walsh era, expediting a switch.  Phil Simms was a sack taking machine who lost his job as a starter in, you guessed it, 1991.  Jim McMahon’s decline was more injury related, but as the guard changed, so did the structure of league-wide passing statistics.

Risk-averse players like early-career Rich Gannon, Erik Kramer, Scott Mitchell, Jeff Hostetler, and Neil O’Donnell were the most employable kind of quarterback of these seasons as passing in the NFL underwent a paradigm shift.  Other players like Warren Moon, Steve Deberg, and Randall Cunningham re-invented themselves and re-emerged as starters as offenses started to take off again in the mid/late 90’s.  Budding stars such as Young, Aikman, and Favre had no such issue beating the shift in the NFL (Aikman might not have had much of a career a decade earlier) and excelling in the new environment.  Passing wasn’t any harder, but point quotas weren’t all the rage anymore.  Teams were out to prove that passing to win was viable whereas in the 80’s, passing was all anyone did to try to catch the best, most balanced teams (Giants and Bears).

Unreal quarterback play bumped point totals for the 94 season, but even after that, defenses were helpless to slow down passing games which eventually (by the late 90’s) drove offensive yardage inflation back towards levels last seen in the 1980’s.  It’s no shock though that in an era defined by the risk-averse behavior of NFL teams, some of the great NFC powers (Redskins, 49ers, Cowboys) just dominated this world and every other team (for example, the American Football Conference) were just kind of there.

2001

Fact: in the era that gave us the 1985 Bears, scoring was inflated.  Also fact: the era that gave us the Greatest Show on Turf oddly depressed scoring.  Awesome.  This is why I love numbers.

Scoring has been higher in the last decade than at any point in the history of football since World War II when the NFL had 8 teams, and if two of them were 40 PPG teams, then you had an inflated league.  2001, however, didn’t get the memo.  In 16 games, the St. Louis Rams scored 503 points.  That’s a lot of points.  It’s deflated, however, if you consider that in 2000, St. Louis scored 540 points after scoring 526 in 1999.  Scoring this year was down about a half point per game.  Unlike the early nineties, the culprit here is probably not rampant conservatism.  And whatever the true problem was, it ended when the Texans were introduced into the NFL shark pool and started hemmorhaging points at a rate that hasn’t actually slowed since the team was created.

This specific year, 2001, seems like more of a fluke than anything.  Why were points depressed?  Well, one reasons was that returners had an incredibly limited impact this year.  This was just a bad year for returning punts especially.  Dante Hall would soon take the league by storm, but in 2001, if Troy Brown wasn’t doing it, no one cared.  Passing numbers weren’t down at all, they were actually up a small tick.  Rushing TD’s fell WAY down though, and first downs by defensive penalties also fell.  Defenses did a lot of the scoring this year as well, if only to make up for the suck of the red zone rushing attacks and special teams.  Special teams also sucked in one other way in 2001: kickers attempted 42 more field goals in 2001 than in 2000, no doubt a result of fewer rushing TDs.  They made just one more field goal in 2001 than in 2000, a remarkable feat.

Absolutely none of these point-depressing effects seemed sustainable, and none did sustain into 2002.  Well, except the kickers were still bad, but scoring soared in spite of them.

2009

2009 was a really high scoring year paced by the ridiculous off the chart offenses of the Saints and Colts.  It’s just become clear that for whatever reason, scoring and specifically scoring via the pass, has taken off since 2007 to the point where we are in uncharted NFL waters with the strength of offense.  And yet, 2009 just doesn’t fit neatly into the narrative.  The average passers and receivers were significantly limited in this one year compared to 2007, 2008, and 2010.  Especially 2008 and 2010.  Which means that some event that happened in 2009 dropped passing offense league-wide, and that corrected itself before 2010.

Before I get to 2009, I’m going to take a moment to explain why 2008 was the strongest offensive year anyone had ever seen at that time.  For one thing, quarterbacks have actually, for the first time in league history, started to adjust to getting sacks.  This was almost certainly a Manning family trend to begin with, but in the last four years alone, a select few quarterbacks have actually posted 16 game season sack rates below 2%.  In 2009, Peyton Manning was sacked just ten times.  Maybe not as impressive as Dan Marino’s 6-sack 1988 when he LED THE NFL IN PASSING ATTEMPTS, but it’s not just Peyton: Eli Manning is consistently under 3%, and Jay Cutler was at 1.5% in 2008.  It used to be relatively common to sack a QB, but the leaguewide rate now sits at 6%, and has been there consistently since 2007.

Turnovers dropped to near unsustainable rates in the last couple of years, eventually settling in at about a 3% interception rate league-wide.

Now, passing statistics have actually increased every year, culminating in this past season, the most efficient year of offensive football ever in 2010.  However, those same indicators would not have predicted a drop in scoring from 2008 to 2009.  The only thing that changed is the rate of interceptions thrown bumped up a fraction in 2009 without a proportional increase in thrown touchdowns (though those also increased, they just aren’t of equal weights).

In 2009, teams actually scored one more TD than they did in 2008.  So you could argue that the real point environment was identical in each of the last three years.  But 288 more points were scored in 2008 than 2009.  Here’s the reasons I collected for that:

• Occurrences of safeties fell
• Teams missed more extra points and converted fewer two point conversion opportunities
• Field goal kicking

Again, the kickers have a huge effect on the number of the points scored per game, although the sample size of a full NFL season is so large that rarely do kickers deviate from expected means, but this definitely happened in 2009.  There was a clear trend in terms of missing extra points because in a sample of 1,100+, the accuracy percentage of this closed skill went from 99.5% to 98.3%.  That’s a really big decline for a nearly automatic occurrence.  Kickers also missed a higher percentage of field goals in 2009 compared to other recent years, but I’m not so interested in field goal percentage decline since I don’t have the average distance of those field goals.  What I am interested in is why there were 70 more attempts in 2008 than in 2009, and more in 2008 than in all other years.

It wasn’t a more aggressive strategy in terms of attempting longer field goals, though the high success rate on 50+ yard attempts definitely explains the differences in FG%.  The attempts decrease from 2008 to 2009 actually came in the 30-49 range, where most field goal attempts occur.  The increased turnovers in 2009 would not have been a cause of increased FGs, if anything, that should have been a limiting effect or the cause of more TDs.

Best I can tell, the reason for higher FG attempt totals in 2008 vs. 2009 is the reason for increased scoring overall in the last four years: coaches are going for it on fourth down.  It just so happened that while this overall trend is paying off with all sorts of increased point scoring benefits in the last four years, it happened to cause a downtick in PPG in 2009.  NFL teams scored 15 more offensive touchdowns in 2009 than they did in 2008 with the passing and rushing environments basically the same.  But they attempted 70 fewer field goals to get those 15 TDs, and they turned the ball over more.  So in this rare, isolated case of the 2009 season, more aggression was worth fewer points scored.  Maybe defenses had information on offense that led to fourth down stops.  But it was probably just a sample size fluke.

Points are up league-wide because of more efficient passing combined with aggressive fourth down behavior by coaches.  It just happened that between struggling kickers, decreased non-offensive touchdowns, and fewer field goal attempts, teams weren’t able to offset these scoring depressing elements with offensive touchdowns.  By 2010, they were, and then some.  We’re in a golden era for scoring points.  In the coming weeks, I aim to take a look at defenses, and whether there is anything they can do to slow the onslaught of points.

FNQB: Are Changes in NYPA a Strong Measure of the Strength of Passing Environments?

Not to possibly curb your desire to read the rest of this article, but the answer to the headline question is “not as strongly as I had hoped.”  The reasoning behind such a conclusion is that my methodology involved using only the passing performances that qualified at a value between 99 and 101 on the pro-football-reference passing index for net yards per attempt, and the result of the research was ultimately a sample size issue.  NYPA is well known as a solid predictive measure of quarterback/offensive aptitude, and so by taking players who were almost perfectly average in this measure between 1970 and 2010, I wanted to see if we could get a good context for the strength of passing in the NFL over the last 40 years.  Still, I have plenty of interesting observations to offer.

The main problem is sample size earlier than 1970.  I had roughly assumed that in any given season, one out of the 25ish passers who threw a majority of their teams passes would end up roughly at league average in YPA.  The accuracy of my methodology makes it back to roughly 1977.  This is a bit problematic, because the passing environment very famously loosened in 1978 with the enactment of the “Mel Blount rules” for cornerbacks, rules that were reemphasized in 2004 in a movement lead by the Colts.  Obviously, I’d like a better sample from 1977 and before, but the “perfectly average in NYPA” method skimps here.

What we can tell about the pre-Blount rules era is best captured by completion percentage, a value that hardly ever exceeded 53% amongst league average passers in the mid seventies.  Although the baseline today is now around 62%, exceeding 60% in the sixties and seventies meant your name was “Tarkenton,” or “Jurgensen,”  or “Anderson.”  Passers who, in a couple of words, were clearly ahead of their time.  After 1975, that began to relax a little bit, and great teams started to have high efficiency offenses, so you’ll see guys like Terry Bradshaw and Ken Stabler with some excellent seasons in the seventies.

If we move our focus to a 16 game NFL season beginning in 1978 with relaxed passing rules, it’s just not clear that the average QB season has resulted in more recent names doing the most to throw their teams to victory.  Take a look at the list of average NYPA seasons re-sorted by yards per attempt efficiency.  David Garrard last season ranks no. 1: the best yards per attempt in the last forty years by a player with an average season.  He is followed by a number of players from the mid eighties: a couple of Neil Lomax seasons, a Warren Moon year, and a Ron Jaworski season.  Then we end up in the middle of this decade.  Anyway, sorting by yards per attempt puts the results all over the map, and doesn’t really do anything to define an era.  The variability of sacks by player doesn’t fall neatly along era lines.  And that is why sorting by yards/attempt efficiency doesn’t work.

Actually, this is where passer rating can really help.  Here is the same list sorted by passer rating, a metric that doesn’t involve sack rate in any way, but does a good job capturing the increases in the efficiency of NFL passing games thanks to declining interception rates.  One of the main observations of this project is that, since 2007, it’s quite common for a player with an average showing in NYPA to put up an 80.0 or better in quarterback rating.  Prior to 2007, it was relatively uncommon for a player’s QB rating to exceed 80 if his NYPA was average.  Since then, it’s happened 10 times.

Rating also shows us that, while there was a measurable late seventies bump in passing offense, it was short lived.  Passing totals REALLY opened up in the mid-eighties, and the truth is (as YPA shows above), NFL defenses were vastly behind for a period of four or five years.  There was probably no larger offense-defense discrepancy (at least in favor of offense) than in the mid-eighties when the Chicago Bears could stop anyone they wanted and every other team could score.  QB rating shows us that from 1983 though 1988, the average QB season came in steadily above the 75.0 passer rating line for the first time with the average QB completing about 58% of his passes.  Those numbers look light today, but defenses are actually a bit better at preventing yards now than they were in the eighties.  It was not at all uncommon to average over 200 passing yards per game for the entire season.

I don’t really know why offensive passing numbers stagnated in the early nineties.  If I had to venture a guess, I would suggest that the offensive-defensive discrepancy seen in the mid-eighties closed, perhaps of factors and adjustments that we would suggest to be rather normal.  I do know that 1994 was an aberration year because a bunch of hall of fame quarterbacks had career years, pushing the average passing year towards the levels we currently see in the NFL.  Following the passing explosion in 94-96, offensive passing totals were pretty stagnant between 1997-2002.

The NFL has been a decidedly different game ever since totals bumped between 2003-2004 and then similarly in 2007-2008.  For the first time since the eighties, we can accept seven yards per attempt to be the standard for passing.  But even then, passing wasn’t necessarily efficient.  It was simply accepted in the 80s that 4ish percent of a QBs passes were going to get picked.  As it became wise to start trading some raw yardage efficiency to take fewer negative plays (sacks, picks), defenses caught up.  But teams got really good at maximizing sack and interception efficiency in the mid nineties to early aughts to the point where passing became an equal or superior option to running the football league wide.  And as the quality of the talent pool saturated with passing and receiving talent after the 2002 season

Back to the top, NYPA isn’t a great measure of era strength, perhaps because it’s too predictive of a measure.  Young players who did average in NYPA typically played a lot more above average years.  Old players who were average in NYPA were near the end of their careers.  It wasn’t easily sortable to the point where the raw number gave a great snapshot of how passers were doing in that season.  I will point out that Adjusted Net Yards Per Attempt correlates almost perfectly with the time horizon examined in this article.

Everything I learned here suggests that passing cycles are cyclical: predictive efficiency (NYPA) ended up as not a great measure of era efficiency because of sample size issues.  Adjusting for interceptions and touchdowns fixes the variance in the sample and shows a long-term linear trend in passing efficiency.  That’s fascinating because it suggests that as offenses have gotten smarter over the years, the defenses have been unable to force offenses (as a group) to negatively trade off points or yards.  For example: it suggests that defenses are helpless to try to swing offensive efficiency in the other direction by increasing sacks or interceptions or passing touchdowns at the cost of yards against.

And despite this, NYPA was identical in 1979 as it was in the 2000s.  The Blount rules had a huge, unbreakable effect on average NYPA, broken only by the unprecedented efficiency in passing we are seeing in the NFL since 2008.  Overall leaguewide passing efficiency, however, started a steady climb then and was undeterred on a league level throughout fluctuations in yard and points expectancy.  Scoring totals, keep in mind, correlate almost perfectly with passing YPA because increased INTs and Sacks (things not captured in YPA) don’t discourage scoring overall, they encourage scores by the defense.

The passing era of the modern (since 2007) NFL is historically unprecedented.  Offenses have never been more efficient, and defenses have never had less of an effect on overall offensive efficiency.  We’re now seeing 30-year long offensive efficiency rates begin to break in the offenses favor.  This is exciting stuff if you happen to be a football history buff.  Despite the sky high offensive efficiency rates, scoring totals will never match the levels of the 1950’s and 1960’s because offenses are now able to keep defensive scoring to a historical minimum.  Despite that, teams are now scoring as much as they did in the 1980’s, when the passing environment was fertile, but also led to high turnover totals and short fields for the other team.  Net yards per passing attempt really didn’t predict any of this: it said pretty much the same thing about offensive power from 1978 through 2006: 5.8 or 5.9 net yards a play.  In the future, however, it might be the best indicator we will have to go on.

All statistics used in this article are courtesy of profootballreference.com