Eight weeks ago, the Vikings were undefeated and ranked No. 1 in DVOA. Now they're 6-6 and probably done. What went wrong?
17 Mar 2005
Guest Column by Michael Horn
"...the stricter rules enforcement... was one of the central story lines of this season,...[and] changed the way the sport was played... Scoring and passing yards per game were up league-wide,... passing yards per game rose by about 20. League-wide, quarterbacks had the highest completion percentage (59.8) and the highest passer rating (82.8) in NFL history." -- Mark Maske, Washington Post, January 11, 2005
From this statement, and many others like it, you would think that this season was a complete pass-wacky free-for-all that saw teams league-wide take advantage of the increased enforcement of the five-yard contact rule. In fact, the changes in the passing game this year were more subtle than that and far from uniform across the league. And the big change was not the increased completion percentage but the role of the tight end.
To begin with, the amount of passing was down this year, not up. There were fewer passes thrown in 2004 than in 2003. Maybe this is because the passes that were thrown were more effective due to the better rule enforcement: the league's passer rating was up 4.5 points.
But pass attempts were also down in 2003 from 2002, so that trend (if it can be called that), started BEFORE the increased emphasis on penalizing illegal contact. Meanwhile passer ratings fell in 2003 along with the decrease in passes, so there does not appear to be a strong correlation between increases in passing effectiveness (at least as measured by the passer rating) and decreases in pass attempts.
Alright, but passer ratings DID jump this year -- a lot from 2003, and some from 2002 -- what caused that? Passer rating consists of four elements: completion percentage; yards per attempt (YPA); percentage of touchdowns per attempt; and percentage of interceptions per attempt. This chart breaks the league-wide passer ratings for 2001-2004 into those components:
|Year||Comp. Pct.||YPA||TD Pct.||INT Pct.||Rating|
The jump in passer ratings was due primarily to higher rating for yards per attempt and a greater percentage of attempts thrown for touchdowns: those are the two components that have increased over the two year period. Completion percentage was up this year, but it was up in 2002 as well, before the illegal contact changes. Interceptions per pass attempt weren't any lower this year than in past years, and in fact were at their lowest in 2002.
The league averaged roughly 6.75 yards per attempt from 2001-2003, but 7.05 yards per attempt in 2004. That's not a huge difference, but when considered over 16,000+ attempts it tends to add up. By the way, there were more passing yards gained in 2002 than in 2004 because there were so many more attempts then. The NFL was more pass-wacky in 2002 than this year, just not as good at it. There were two reasons why YPA were up in 2004. First, wide receivers were gaining more per catch compared to the last three years:
Since that effect was primarily seen in 2004, it MAY have something to do with the increased enforcement of the rules. Running back and tight end yards per catch have stayed roughly constant over this period. The second reason that YPA was up over the last two years is that passes to running backs went down. This was true in both in raw numbers of passes thrown to them (a.k.a. "pass targets") and in the percentage of passes thrown to them compared to passes thrown to wide receivers and tight ends. While the "market share" of wide receivers has stayed about the same during this period, tight ends were seeing a fairly significant increase in the percentage of passes targeted at them.
This growth in tight end "market share" has been steady for years now, so it was not just due to the 2004 rule enforcement emphasis. (Note: This table does not include passes with no intended receiver or passes to other positions.) What does the percentage of passes thrown to running backs have to do with yards per attempt? Simple: running backs catch shorter passes than tight ends. Therefore, when quarterbacks throw more to their tight ends than to their backs, YPA tends to go up. This is true even though a pass to a running back is a higher percentage throw (73% completed) than one to a tight end (63%): the extra yardage gained more than offsets the lower percentage pass.
So, to summarize, YPA (and passer ratings) have increased from 2001-2004 because: 1) wide receivers were gaining more with each catch they made and 2) for three straight years, teams have thrown more often to their tight ends than the year before, and less often to their running backs.
The second component of the league-wide passer rating that went up in 2004 was touchdowns as a percentage of attempts:
Since this number is pretty much the same from 2001-2003, it is possible that here is another place where the increased rule enforcement might have had an impact. And although the overall change is small in percentage terms, as we saw earlier it has had even more of an impact on the passer rating than the increased YPA. And which position has most contributed to the improvement?
|Position||TD||TD Pct.||TD/Target||TD||TD Pct.||TD/Target||TD||TD Pct.||TD/Target||TD||TD Pct.||TD/Target|
The big change has been at tight end, again. Last year, compared to 2002 and 2003, a higher percentage of passing touchdowns were caught by tight ends, and there was a big jump in the percentage of passes targeted to tight ends that were caught for touchdowns. Now, the fact that these two statistics were at a similar level in 2001 might make you think that 2002 and 2003 were just historical anomalies. But actually, 2001 and 2004 are the historical anomalies. The 26% of passing touchdowns thrown to tight ends in 2004 is actually the highest percentage in the history of professional football, at least since the beginning of the AFL in 1960. And with the exception of 2001 and 2004, tight ends had not caught at least 21% of passing touchdowns since 1985. Combine the rise in passes to tight ends with the jump in the percentage of those passes that go for touchdowns in 2004, and you have an explanation for the rise in touchdowns per pass attempt.
Interestingly, tight ends did not see a big increase in passes thrown to them at the goal line (defined here as inside the five-yard line), either in total targets or in their share of goal line targets. They did, however, see a jump in passes thrown to them in the red zone, and this increase came at the expense of the wide receivers. This table is just 2002-2004 for space considerations:
|2002 Pass Targets||2003 Pass Targets||2004 Pass Targets|
|Position||Targets||Red Zone||% of RZ||Goal Line||% of GL||Targets||Red Zone||% of RZ||Goal Line||% of GL||Targets||Red Zone||% of RZ||Goal Line||% of GL|
It's possible that the changes to tight end use, in the red zone and overall, are partly due to the increased enforcement of the illegal contact rule. But I think it has more to do with the talents of individual players and some key coaching changes.
As some of these changes took place over a two period, the increased rule enforcement this year is not the sole explanation for what has been going on. "The tight end" is not the only answer either, but it's the biggest part of the answer.
There has not been a league-wide surge in the use of tight ends, as you might expect if the game had changed solely due to a rule change that everyone could see and take advantage of. In fact, from 2003 to 2004, 17 teams actually saw the share of passes targeted to tight ends decrease, even though the overall movement was in the other direction. From 2002 to 2004, the trend toward tight ends was stronger: they increased their share on 23 teams while losing share on nine. Because of this mixed movement, with some teams using their tight ends more, and some less, it's very hard to isolate a specific cause: the trend is more the net accumulation of a number of changes. Here's several cases that exemplify the overall effect:
|Share of Passes Thrown to Tight Ends by Year|
|Team||2002||2003||2004||Reason for Increase|
The first two cases are easy to explain, as (alleged) offenses that hardly ever used the tight end left the league with Dick LeBeau after the 2002 season and Steve Spurrier the following year. They were replaced by coaches that used their tight ends an average amount (note I'm counting Chris Cooley as a tight end even though he often plays the hybrid H-back position). Those two coaching changes alone added 92 more targets to the league-wide tight end ledger from 2002 to 2004, accounting for about a quarter of the shift in target share towards tight ends.
The third case (SD) is another easy one: Antonio Gates. A change in how Marty runs his offense is a sign of either the apocalypse or a really good player. Maybe there's some cool Brees effect in there, but Gates is an example of an individual so talented that his coach and quarterback had to throw him the ball. That turned the Chargers into the most tight end friendly offense in the league.
The next situation (KC) is a bit more complicated. Sure, Tony Gonzalez is great, but he's had the same quarterback and coaching staff all three years so you wouldn't expect a huge change in his usage in that time. For the first year and a half of it, he saw six targets per game. Then in the middle of 2003, for reasons not entirely clear (except that it wasn't increased rules enforcement), he began seeing an additional couple of passes each week until mid-2004. Then Priest Holmes got hurt and Gonzalez saw another pass each week -- until the final game of 2004, when he was targeted 21 times in a successful (but blatant) effort to get him the record for most passes caught by a tight end in a season.
|Average Passes per Game
to Tony Gonzalez
|2002 through Game 8, 2003||6.0|
|Game 9, 2003 through Game 8, 2004||7.9|
|Game 9, 2004 through Game 15, 2004||8.9|
|Game 16, 2004||21.0|
In 2002, KC threw 106 passes to its tight ends. In 2004, it threw 172. The increased use of Gonzalez in 2004 (compared to 2002) added 47. This still leaves around 20 additional tight end targets in the Chiefs offense to account for:
|Average Passes per Game
to Jason Dunn
Why did the Chiefs suddenly start throwing the ball to Jason Dunn more in 2004? Here's my answer: the Chiefs drafted a tight end, Kris Wilson, in the 2nd round. Wilson looked great in pre-season, but this was seen by the media and fans as a strange move because the Chiefs already had a star tight end and gaping black holes all over their defensive universe. Wanting to take advantage of the rookie's abilities, and needing to justify the pick, the coaching staff put into the offense a number of plays using two tight ends.
Then Wilson broke a fibula just before the season started (the Chiefs thought enough of him, though, that they didn't put him on IR). But now they had all these great plays for two-TE formations. So Dunn saw the ball a lot more. Either that, or the Chiefs discovered that Jason Dunn was uniquely suited to taking advantage of the increased enforcement of the illegal contact rule. By the way, in the Dawg Pound there was a comparable situation where the drafting of Kellen Winslow v2.0 led to increased use of tight ends even after he got hurt: the tight end share of pass attempts went up 12% in Cleveland this past year.
So the shift in the Chiefs' passing offense towards the tight end over the last two years can be explained by the presence of a talented player at the position (Gonzalez); the addition of more talent there (Wilson, with Dunn standing in); and the loss of talent at another position (Holmes).
That bring us to Tampa Bay, another example where one player made a big difference in how a team used its tight end. From the beginning of 2002 through week 4 of 2004, Jon Gruden's Bucs ran one of the least TE-friendly offenses in the league. In those 36 weeks, with Ken Dilger as the main man at the position, the Bucs threw 11% of their passes to their tight ends. About a third of the time, the tight ends saw two or fewer targets, about a third of the time it was three or four targets, and the remaining third it was five or more. (In one extreme week, Dilger saw 11 targets when Tampa played his former team, the Colts; maybe they should have spent more time killing the clock). Then for 11 weeks in 2004, tight ends started seeing 18% of the Bucs passes, with only one game where they saw two or less, and eight weeks where it was five or more. Dilger remained the tight end during this span. And he was still there when in week 17 when tight end targets dropped back to three. What happened?
|Tampa Bay Passes to Tight End 2002-2004|
|2 or fewer
|5 or more
|Griese as QB||11||1||8||5.5|
|All Other QBs||37||12||12||3.8|
Brian Griese became the starter in that 11-game span. He had learned the pro game in a very tight end-friendly offense (Denver) and that apparently carried over into how he ran the offense in Tampa. It had a fairly significant effect in making this team about average in throwing to the tight end. It's not clear if that's the long-term direction that Gruden wants this offense to take, but for the short-term, Griese took it down that path.
These five cases typify what happened across the league when teams increased their use of the tight end from 2002-2004: coaches changed or personnel changed. The strongest personnel effect could be seen when a talented player was added at tight end, but changes at other positions could make a difference too.
Similarly, where teams saw a decrease in the use of their tight ends, it can be traced back to coaching and personnel changes. Again, five examples:
|Share of Passes Thrown to Tight Ends by Year|
|Team||2002||2003||2004||Reason for Decrease|
|NYG||31%||19%||25%||Injury, new coach|
The Jaguars and Cardinals saw drops in their tight end usage when Del Rio and Green arrived with offenses that looked to throw elsewhere. Prior to their arrival, these teams had moderately TE-friendly passing attacks. With Del Rio, this change was not foreseeable, but Denny Green had run a wide receiver-oriented passing game in Minnesota that did not use the tight end very much.
Next case: in 2002, Fassel's Giants threw more to its tight ends than any other club except the Ravens. With Shockey out for several weeks in 2003, the Giants dropped to near-average and Fassel dropped to consulting. In comes Coughlin, Shockey gets healthy, and the tight end share of the Giants' passes goes up -- but not to 2002 levels. A statistically insignificant but interesting fact: Coughlin's 2002 Jaguars and 2004 Giants both threw to their tight ends a quarter of the time -- despite an apparent disparity in talent at the position when comparing the two teams (Kyle Brady/Pete Mitchell vs. Shockey). More evidence of Coughlin's flexible methods?
The fourth example is straightforward: Todd Heap got hurt. Without him, the Ravens passing game (I use the term loosely here) looked elsewhere. Can you even name the other Ravens tight ends?
This brings us to Houston, the most complicated case of the disappearing tight end pass share. In 2002, the share of the Texans passes thrown to the tight end was the sixth-highest in the league. The next year, it was middle of the pack. And in 2004, it was the third-lowest, with only the consistently TE-unfriendly offenses in St. Louis and Pittsburgh using their tight ends less.
Here's what I think happened. In 2002, the expansion Texans had one of the worst offenses in modern NFL history, anchored (as in dragged to the bottom) by a horrendous offensive line that got its quarterback sacked a league-record 76 times. Rookie quarterback David Carr threw to his top wideout, Corey Bradford, 106 times; to his #2 receiver, Jabar Gaffney, 80 times; and to his other wide receivers only 64 times total. One conclusion: he did not have a lot of time to go to secondary reads and therefore went to his tight end a lot as his safety valve (91 total targets). He did this more often than he threw to his running backs (79 targets).
Fast forward to 2004: Texans have improved their offensive line to merely bad, Carr has more experience, and two key offensive weapons are added: Andre Johnson and Domanick Davis. Davis alone has more passes thrown to him (85) than the whole running back corps in 2002. Johnson sees 138 passes. Gaffney is still the #2 wide receiver target with 68 targets. And the rest of the wideouts see their workload go up to 149, well above their 2002 total. With a combination of experience and better protection, it appears Carr has more time to look for his wide receivers, while ensuring the primary weapons in the passing game (Johnson and Davis) get most of the load.
The odd man out is tight end Billy Miller. He went from second on the team with 81 targets in 2002 to seventh with 34 targets this past season. The #2 tight end in Houston in 2004 was Mark Breuner, who was brought in from Pittsburgh after the 2003 season. It is interesting that a back-up tight end with a long history as a blocking (versus receiving) option in a TE-unfriendly offense was added in the off-season. Clearly, at least in retrospect, the Texans did not expect to use their tight ends much in the passing game this past season. A combination of coaching decisions and talent changes resulted in a decreased use of the tight end in Houston.
Again, these five cases typify why some teams used their tight ends less this past season than in previous years: coaches changed and available talent changed.
A good part of my analysis has argued that increased talent among tight ends has contributed to their greater use in the passing game (or maybe I should say different talent, since blocking is a skill just as pass-catching and route-running are). To make this possible would require an influx of tight ends through the draft (leaving aside undrafted free agents like Antonio Gates). In fact, the last few years have seen tight ends drafted earlier and, in one year, much more often. Since the league went to a seven-round draft in 1993, the number of tight ends drafted, both total and as a percent of picks made, has remained remarkably stable, with one exception:
|Tight Ends and the Draft|
|Total TE Drafted||13||13||13||15||15||13||12||12||14||24||13||14|
|% Picks Used on TE||5.8%||5.9%||5.2%||5.9%||6.3%||5.4%||4.7%||4.7%||5.7%||9.2%||5.0%||5.5%|
Two years ago, there was a surge in the number of tight ends drafted. And it wasn't just a surge in the numbers, it was also an increase in the presumed quality, as there was a jump that year in the percentage of first round picks used on tight ends. As many tight ends have been taken in the first round over the past three seasons as were taken in the first round between 1993 and 1999:
|First Round Tight End Picks|
|First Round Picks on TE||1||0||2||1||2||0||0||2||1||3||1||2|
Meanwhile, there has also been an increase in the number of tight ends chosen on the first day (rounds 1-3). Again, there were quite a few tight ends drafted on the first day in the mid-90s, then a decline. But here, the increase doesn't show up until 2003. It's only the last two years that have seen a large increase of tight ends drafted early.
|Percent of Tight End Picks Made in Rounds 1-3 (over 40% in bold)|
|% Picks Used on TE||30.8%||15.4%||46.2%||46.7%||40.0%||23.1%||33.3%||25.0%||28.6%||25.0%||46.2%||42.9%|
Together, these numbers indicate a gradual increase in the amount of talent coming into the league at the tight end position over the last few years. In addition, more of that talent is sticking around.
|Number of Tight Ends Drafted In The First Round On Active Rosters|
|First Round TE on Active Rosters||5||5||7||8||8||8||7||7||8||10||11||13|
There has been a steady increase in the number of blue chip tight ends (at least as defined by draft position) that are playing in the league, to what is almost certainly an all-time high (only a small part of which is accounted for by the increase from 28 to 32 teams in this period). This accumulation of talent has then been reflected in the increased use of tight ends in the passing game.
There has not been a one-year, league-wide surge in the use of tight ends, as you might expect if the referees' enforcement of the rules was the sole reason for the change. The game has evolved as it always has, because new ideas, coaches, and talents come into the NFL along with different rules (or their enforcement). The four years looked at here may not be enough time to assess these changes. But in this short time, the league, as a whole, has put much more emphasis on tight ends as measured by how they are drafted and used. This did not start with the increased enforcement of the illegal contact rule, and won't necessarily end if that emphasis fades.
Michael Horn, a retired Army officer, is a graduate of a small engineering college on the Hudson that once ranked 299th out of 300 on a list of party schools. Not content with the rigor of that education, he later attended #300, the University of Chicago, for a Masters in European History. He now works as a defense (not football, the other kind) analyst in the military-industrial complex and occasionally comments as MRH on Football Outsiders discussions. An earlier version of this article originally appeared on the web site fantasybeat.com, where Mike writes an "Inside the Stats" feature. If you have an idea for a guest column, something that analyzes the NFL from a distinctive point of view, please email us at firstname.lastname@example.org.
1 comment, Last at 07 Feb 2006, 4:31am by adam