During the offseason, I’ll be looking back at certain aspects of the Jets’ season by analyzing data compiled from all nineteen games, rather than watching film. I will be tackling as many diverse topics as possible, but welcome your suggestions or requests in the comments.
This week’s BGA focuses on Dustin Keller. I’m going to be analyzing how well he performed when lined up as a wide receiver and compare this with his work at the tight end position.
I’m then going to try and figure out the reasons behind any trends that are apparent. Ultimately, I’ll be giving consideration to whether lining up as a wideout is something he should be doing more or less of.
Once again, this article has used data exclusively provided to us from the guys at PFF. Our thanks, as ever, go out to them.
When the Jets traded up for Dustin Keller in 2008, it wasn’t because they needed a tight end. At that time, their biggest need was a pass catcher and, in a weak year for wide receivers, Keller was the best available pass catcher and the first one to come off the board.
Keller’s athletic ability made him an option as a slot receiver, or occasionally even split out wide, which was sure to be a boost for a team that began the season with Jerricho Cotchery and Laveranues Coles as its starters. Although Coles left after 2008, the need for Keller to play as a receiver dissipated somewhat when they traded for Braylon Edwards a month into the 2009 season and then almost completely when they added Santonio Holmes before the 2010 season. However, Keller has still seen plenty of action as a wideout, where there have been injuries or when the Jets have gone to a spread offense.
The first three years of Keller’s career have been somewhat frustrating, despite some tantalizingly promising signs at various points. Although his numbers did not change much from year one to year two, that can be attributed to the fact that the Jets passed a lot less. As a percentage of the total passing yards, there was a substantial increase. In his third season, he did make some strides in terms of his productivity, but his output was ultimately lower than might have been expected following a fast start.
The question therefore becomes: How productive was he when he lined up as a wide receiver? Did this make him more of a threat or was it something that served to take him out of games?
For the purposes of the article, I will separate the data from when he was split out wide from the data where he was lined up in the slot. Note: To be treated as being in the slot, it is irrelevant whether Keller was in a three-point stance. Instead, what I am looking for is that he lines up outside the defensive end’s outside shoulder and is therefore not in a position to block him.
Let’s start off by looking at how often Keller was lined up as a wide receiver over the course of the season. This may surprise you.
- Slot receiver: 241 snaps (25% of total)
- Split out wide: 48 snaps (4% of total)
If we just look at snaps where Keller ran a route on a pass play, these percentages increase.
- Slot receiver: 212 snaps (38% of all routes run)
- Split out wide: 37 snaps (7% of all routes run)
Already we can see that on pass plays where he was a potential target, the number of snaps he had as a receiver was almost half of the time. It will be interesting to see if that accounts for almost half of his targets.
Just to show how worthwhile it is to consider how he performs as a wide receiver, he actually ran almost exactly twice as many routes from the wide receiver position as the Jets’ number four receiver, Brad Smith. Smith saw a total of 124 plays where he ran a route on a pass play, 56 from the slot. Again it will be interesting to see if Keller was targeted twice as often.
The next thing we need to consider is how those snaps broke down over the course of the season. With no Santonio Holmes in the lineup for the first four weeks, you might expect him to see more action at wide receiver. Similarly, Jerricho Cotchery suffered an injury in Week 10 against the Browns and missed a few games and Brad Smith was hobbled during the postseason.
Keller was used as a wide receiver for 12 plays per game on average for the first three weeks. That dropped to three in Week 4, but the Jets were in a lot of heavy formations, as they ran the ball a lot to exploit the Bills weak front seven. When Holmes returned, they used him less, as you’d expect. Keller was used just seven times at wide receiver in the two games leading up to the bye week.
After the bye week, there was a sudden surge in how often he was used as a receiver (24 times per game for the next three weeks). This, we can attribute to the fact that the Jets had to use a lot of hurry-up in each of these games, as they were either losing, or trying to score at the end of the game.
Following Cotchery’s injury, Keller ran WR routes on nine snaps in Week 11 (another late-game comeback) and 17 in Week 12. Cotchery returned for Week 13, but Keller saw plenty of action at wideout over the next two weeks (51 snaps), as again the Jets trailed more or less throughout in both games.
In Weeks 15 and 16, his WR snaps dropped back down to 17 and 16, respectively, but this was still higher than when Holmes was unavailable at the start of the year. He did not play WR in Week 17.
In the postseason, with Brad Smith not 100%, Keller ran WR routes on 13, 19 and 21 snaps over the three games. Again this was more than at the start of the year.
Looking at Keller’s numbers for the year, he got off to a really fast start over the first four games, but once Santonio Holmes returned, his numbers started to fall off. Part of that is, of course, attributable to the fact he saw fewer targets and less playing time overall, but I did notice teams doing a good job of gameplanning against Keller too, whether that be by jamming him at the line, or getting a safety to come up and run with him.
After the bye week, part of the increase in his usage as a wideout was attributable to the Jets having to use more spread formations because they were trailing or running a two minute offense. However, he was still seeing more action there in games where this was not the case, so I have to conclude that this was a conscious adjustment to try and get him back to being more involved in the offense. Whether this is successful is debatable. He did catch at least three passes in all but two of the games after the bye week, but he only had more than four catches twice. It’s possible that – without this adjustment – he might have disappeared from the offense to an even greater extent.
Keller’s 2010 Productivity
Let’s now consider how the Jets fared when they targeted Keller as a receiver, and compare this to when they threw to Keller the tight end.
Tight End – 44-69-515yds, 4TD, 1 int
Wide Receiver – 25-46-294yds, 1TD, 2 ints
The statline when throwing to Keller as a receiver doesn’t look too bad, but it only translates to a QB rating of 63.0. The completion percentage is 54%, right at Mark Sanchez’ career average. However, there were four incompletions that were negated by a penalty. This would have dropped the figure to 50% exactly.
As a tight end, he’s that much more productive, catching 64%, although the average yards per catch is exactly the same (11.7).
When Keller lines up as a tight end and runs a route on a pass play, he is targeted 24% of the time. As a receiver, he is only targeted 18% of the time, although this increases to 20% if we just consider when he’s in the slot. As noted above, he runs 45% of his routes from the WR position, but his catches from that position only account for 36% of his total.
On the whole, therefore, he was more productive from the tight end position. However, things get more interesting when we break down the game-by-game splits.
When Keller was split out wide, he caught three passes on three attempts, for a 24 yard touchdown in Week 3, and gains of two and four yards in Week 15 and the Divisional Playoff against New England. Although we can momentarily wonder whether that’s something they should have explored more, let’s shelve this data for now and just compare the tight end and slot receiver splits.
Before the Bye Week
- As a slot receiver: 1-6-9yds, 0TD, 1 int
- As a tight end: 22-33-310yds, 5TD, 0 int
There’s a huge disparity here, but only 117 of those yards came after halftime in Week 3, so this was the point at which the Jets had realized that teams were making a concerted effort to stop Keller, so they made the adjustment of using him as a receiver more and, logically, looking elsewhere.
Note: Even the one catch he made was a bad play, because it came on a late fourth and ten against Baltimore and he stepped out before the marker. Maybe this suggests he was still learning the ropes in terms of running routes from the receiver position too, perhaps accounting for the slow start to some extent.
What happened to the breakdown after the Jets adjusted?
Week 8 to Week 13 (inclusive)
- As a slot receiver: 12-16-175yds, 0TD, 0 int
- As a tight end: 6-16-37yds, 0TD, 1 int
All of a sudden, there’s a huge discrepancy in the opposite direction. Clearly the Jets were wise to use Keller more as a receiver because if they’d given his reps to someone like Brad Smith, then he’d have all but disappeared from the gameplan.
Down the stretch, teams were forced to think twice about keying on Keller when he lined up at the tight end position and the Jets sought to attain a more balanced split between the numbers and an increase in production overall.
Week 14 onwards
- As a slot receiver: 9-21-79yds, 0 TD, 1 int
- As a tight end: 18-23-253yds, 0 TD, 0 int
At the end of the year, Keller finally rediscovered how to be efficient and productive from the tight end position. As teams often sought to slow him down by putting a safety on him when he went into the slot, the Jets responded by moving him back into the tight end position from where he could be more effective without a big, physical guy on him.
Hopefully, this demonstrates that moving him into a slot position is an adjustment they can make to prevent a lineman chipping him at the line and slowing his route down. If the opposition counters by putting in more pass-oriented personnel, the Jets can revert to using him as a tight end again.
Comparison with Brad Smith
As noted earlier, Keller ran almost exactly twice as many routes from the receiver positions as Brad Smith did. We can immediately see how much more productive Keller was, because Smith was only targeted eight times, catching six.
This serves as a good indicator that using Keller there rather than a backup receiver is likely to be more fruitful, perhaps due to how the defense will match up with their personnel. There’s obviously an inherent advantage in being able to line up outside when you aren’t a wideout.
Dustin Keller still isn’t the finished article as a tight end. He drops a few too many balls and is still a work-in-progress as a blocker. However, he is versatile enough to contribute in a number of different ways, all of which seem to be forcing the defense to make adjustments, based on what I saw during the season. Even if he isn’t being statistically productive, Keller contributes as a decoy, often running clear-out routes to open up the underneath areas for the likes of LaDainian Tomlinson to exploit.
As closely as I watched the team all season, I had no idea how dramatic the splits were in terms of where he did damage from. Hopefully, that’s a sign he will step it up in multiple areas next year.
If Jeff Cumberland can contribute this season and showcase similar levels of versatility, then this will mean he and Keller are interchangeable, which could make the Jets offense all the more potent.
More BGA next week. Keep those ideas coming!