Writers of Pro Football Prospectus 2008

Most Recent FO Features

CrabtreeMic09.jpg

» Week 6 Quick Reads

The Oakland Raiders as a team, and Amari Cooper in particular, are having a disappointing season. Somehow, though, Michael Crabtree is having a career resurgence. What does it tell us when one wide receiver excels while his teammate struggles?

26 Jul 2017

Jarvis Landry's Value, Part I

by Scott Kacsmar

Now that Football Outsiders Almanac 2017 is available, we can expand on certain topics and statistics that were introduced in this year's book. One of those is the conundrum known as wide receiver Jarvis Landry. I spent a portion of the Miami chapter talking about his actual value to the Dolphins since he will be an unrestricted free agent after this season. So far, negotiations for a contract extension have been quiet as we get to the start of training camp.

Look, in a league where Tavon Austin has signed a four-year extension worth $42 million, it is all but a given that the Dolphins will keep Landry in Miami, as they should. The question is how big of a deal should he get given his unique playing style? Antonio Brown currently has claim to the league's richest deal for a wide receiver at $17 million per season, and players such as A.J. Green ($15 million annually) and Julio Jones ($14.3 million annually) aren't far behind. Fortunately, the wide receiver market doesn't seem to be as crazy as the quarterback market, where there's a thirst to make your guy the highest paid in history for fear of losing him. However, much like Derek Carr compared to his quarterback peers, Landry's performance through three seasons is not on par with the best wideouts in the NFL.

The following is an excerpt from the Miami essay in FOA 2017:

The No. 1 option in this offense still seems to be Jarvis Landry, who is an unusual No. 1 wide receiver since he led all receivers in yards gained from the slot last year. Sure, Julian Edelman serves a similar role for the Patriots, but that offense still features a dominant threat in Rob Gronkowski, and also traded for Brandin Cooks. Landry has been far and away the targets leader in Miami, but that has not necessarily been a good thing for his offense.

For many fans, Landry is even more divisive than his quarterback. While 2016 was his best season yet, Landry’s ranking as the No. 7 wide receiver in the NFL Network’s Top 100 Players of 2017 is a bit absurd. Landry still did not crack the top 40 wide receivers in DVOA, and finished in the top three in failed receptions for the third year in a row. Landry has only averaged a paltry 10.6 yards per reception for his career, though he raised that average to 12.1 yards in 2016. He only catches about a handful of touchdowns per season too, so he doesn’t have specific talent for breaking big plays or producing in the red zone.

The number most associated with praise for Landry is 288, the record number of receptions for a player in his first three seasons. Of course, fellow 2014 rookie Odell Beckham Jr. also has 288 catches, but he has 35 touchdown catches compared to 13 for Landry. The fact is that Landry’s catches are just not as valuable as those made by his peers.

We then looked at correlation between wide receiver stats and passing DVOA, but that is something we wanted to expand on in a two-part study with data for more players than we had space for in the book.

Landry is hardly the first Miami receiver to fit the "catches a lot of short, non-beneficial passes" mold, but I sure don't recall anyone hyping up Davone Bess or Brian Hartline in this fashion. In fact, since 2008, Miami has three of the five wideouts with the lowest touchdown percentage (minimum 250 receptions).

WR: Lowest TD% Since 2008
Rk Player Catches TD TD%
1 Jason Avant 316 10 3.2%
2 Harry Douglas 309 10 3.2%
3 Davone Bess 363 14 3.9%
4 Brian Hartline 344 14 4.1%
5 Jarvis Landry 288 13 4.5%
6 Danny Amendola 365 17 4.7%
7 Julian Edelman 425 24 5.6%
8 Wes Welker 695 41 5.9%
Minimum 250 receptions

Even Wes Welker caught his first break in the slot receiver role with the Dolphins, and his performance was enough to catch the eye of Bill Belichick and the Patriots for a trade in 2007. With that trade, New England basically defined the modern slot receiver, and has successfully cloned Welker with Julian Edelman, one of the latest Super Bowl heroes for New England. The Patriots also had to grab Danny Amendola for good measure, but Edelman has been the team's No. 1 wide receiver since 2013.

Despite Edelman's production, including the two toughest catches in his team's two Super Bowl wins, he doesn't generate the type of buzz that Landry has been receiving early in his career. Maybe it's the draft status -- Landry was a second-round pick compared to seventh round for Edelman -- and likely it is the shadow cast by Tom Brady, Belichick, and the plug-and-play New England machine. But Edelman was only voted as the 13th-best wideout by his peers, if you believe his peers are solely creating those rankings.

The 12th-ranked wideout on that list this year was T.Y. Hilton, who had more receiving DYAR (360) in 2016 than Landry has had in his career (344). It is one kind of crazy to put Landry up there with Brown, Jones, Green, Beckham, and Mike Evans, but what exactly are people seeing that even puts Landry above Edelman, Hilton, Doug Baldwin, Dez Bryant, and Jordy Nelson?

I knew that beyond DYAR (total value) and DVOA (value per play), our data on receiving plus-minus (Landry ranked 16th in 2016) and YAC+ (No. 5) was more kind to Landry's season. Still, the three-year stats don't paint the picture of a top-flight receiver, and the results of the Miami offense also reflect that.

It is no secret that I often fish for ideas and other viewpoints, sometimes regrettably, on Twitter. I wish I could recall and credit the fantasy writer who pointed this out weeks ago, but I saw a table that showed that Miami's offensive production was actually better when Landry was rarely targeted in the passing game. This led me to collect our weekly DVOA numbers for offenses since 2014 to compare to the production (targets, receptions, and receiving yards) of top receivers. Would I find any strong correlations there? Probably not, since the passing game is far too complex to have its success tied to the efforts of one wide receiver, but the results are suggestive that too much of Landry is not a good thing for Miami.

My selection of 28 wide receivers came down to a few requirements. They had to play for the same team (an attempt to mitigate the effects of multiple quarterbacks and offensive casts) in each of the last three years while recording at least 2,000 receiving yards since 2014. Any player who missed an entire season was excluded, so no Jordy Nelson here because of 2015. I only looked at regular-season games, and decided to just exclude games entirely where the receiver did not play rather than give him all zeroes for his production.

The following table shows the correlation by game between team passing offense DVOA and that wide receiver's targets, receptions, and yards. The DVOA is a team rate rather than the individual quarterback DVOA's seen here. The correlations are ranked for receptions and yards, though it's not an absolute value ranking for strength regardless of direction. A negative correlation is a relationship where one variable increases as the other decreases, and vice versa. The player's standard deviation in his targets is also shown.

WR Stat Correlation with Pass Offense DVOA by Game, 2014-2016
Player Team Targets Receptions Rk Yards Rk Target STDEV Rk
Eric Decker NYJ 0.26 0.52 1 0.62 1 2.88 6
Dez Bryant DAL -0.02 0.44 2 0.50 3 2.85 4
Alshon Jeffery CHI 0.03 0.33 3 0.31 8 3.59 21
Mike Evans TB 0.08 0.31 4 0.43 6 3.79 26
DeAndre Hopkins HOU 0.09 0.31 5 0.46 4 3.70 24
Allen Hurns JAC 0.20 0.27 6 0.56 2 2.94 7
Allen Robinson JAC 0.21 0.26 7 0.26 15 3.13 10
Antonio Brown PIT 0.04 0.23 8 0.31 9 3.60 22
Steve Smith BAL 0.03 0.22 9 0.44 5 3.24 17
T.Y. Hilton IND 0.00 0.21 10 0.14 24 3.21 14
Emmanuel Sanders DEN -0.01 0.20 11 0.21 18 3.54 20
John Brown ARI -0.06 0.18 12 0.22 17 3.18 13
Odell Beckham Jr. NYG -0.06 0.17 13 0.30 10 3.74 25
Pierre Garcon WAS 0.05 0.16 14 0.16 22 2.86 5
Player Team Targets Receptions Rk Yards Rk Target STDEV Rk
Julian Edelman NE 0.08 0.13 15 -0.05 28 3.30 18
Jordan Matthews PHI 0.12 0.12 16 0.16 23 3.16 11
Doug Baldwin SEA 0.06 0.12 18 0.37 7 3.02 8
DeSean Jackson WAS 0.09 0.12 17 0.29 12 2.57 2
A.J. Green CIN -0.13 0.11 19 0.20 20 3.98 28
Julio Jones ATL -0.19 0.11 20 0.28 13 3.79 27
Golden Tate DET -0.17 0.10 21 0.08 27 3.22 16
Kenny Britt LARM -0.20 0.10 22 0.18 21 2.80 3
Randall Cobb GB -0.01 0.03 24 0.28 14 3.39 19
Larry Fitzgerald ARI -0.13 0.03 23 0.13 25 3.17 12
Demaryius Thomas DEN -0.12 0.02 26 0.30 11 3.02 9
Brandin Cooks NO -0.23 0.02 25 0.21 19 2.44 1
Sammy Watkins BUF -0.38 -0.12 27 0.25 16 3.22 15
Jarvis Landry MIA -0.26 -0.14 28 0.13 26 3.64 23

Truth be told, there are 28 stories to unpack here, and not enough time to really give each of them justice. As mentioned before, too many variables go into the success of a passing offense to think that there would be strong correlation with just one wide receiver, no matter how good he is perceived to be. If anything, this table shows that quarterbacks deserve more credit for their success instead of getting the "yeah, but he plays with a great receiver" criticism. Let's go through the table piece by piece.

Correlation between targets and passing DVOA

Only Landry (-0.26) and Sammy Watkins (-0.38) were stronger than -0.25 in correlation between targets and DVOA. The fact that half the players had negative correlation with their targets is not that surprising, nor is the fact that there aren't any common characteristics among the negative correlation group or the positive correlation group. Some had a great quarterback, and some didn't. Some had great receiving teammates, and some didn't. Some were more slot/dink-and-dunk oriented, and some were deep threats.

The average correlation should be around zero here given all of the game script possibilities at play. An offense could succeed with its top receiver getting shut down because the quarterback is good at exploiting other matchups, or simply has another great weapon to throw to. MVP Matt Ryan did this several times last season when Julio Jones wasn't a big factor. There can also be games where the quarterback is force-feeding targets to his main receiver with little success, creating a negative correlation with DVOA. On the flip side, some of those days where a top wideout gets a ton of targets could lead to a field day, such as Antonio Brown's 16-of-18 for 189 yards and two touchdowns against the vaunted 2015 Broncos. Also, only getting a few targets could still lead to a huge payoff if the receiver makes them count, like when Landry had three catches on four targets for 108 yards and a touchdown against the Jets last December. Landry has just three career touchdown catches of 15-plus yards, and his longest (66 yards) came in that game.

In the case of Watkins, Buffalo has had three games with a passing DVOA over 100% since 2014. It just so happens that Watkins had no catches on three targets in the 2015 opener against Indianapolis (135.9% DVOA) where Tyrod Taylor was otherwise flawless (14-of-16 for 195 yards, touchdown, and no sacks). As a reminder that we are dealing with small sample sizes ranging from 33 to 48 games, if that one weird game was excluded from the sample, then Watkins' correlation between targets and DVOA would be -0.32 instead of -0.38. Two weeks after Indianapolis, Taylor was again amazing in a game in Miami where Watkins only caught one of his two targets for 39 yards. In the three games where Buffalo had -50.0% DVOA or worse, Watkins had eight-to-nine targets in each game, and generally struggled. However, Watkins has had five 100-yard receiving games in the seven games where Buffalo had at least 60% DVOA, but it's those top and bottom performances that really skew the numbers. Remaining healthy and always available for Taylor is still extremely important for Watkins to achieve in 2017.

The AFC East really gives us more to talk about than any other division here. Brandin Cooks hasn't even officially debuted in it yet, but the former Saint will play for the Patriots this year. He had the third-most negative correlation (-0.23) with his targets and what could be considered Drew Brees' DVOA, but again, that's really a case where a one-game sample is skewing things. Against the Rams last year, the Saints' passing game was deadly (29-of-37 for 360 yards, five touchdowns and 78.9% DVOA), but Cooks didn't get a single target despite playing 45 snaps (five more than Willie Snead and five fewer than rookie Michael Thomas). Cooks never appeared on a 2016 injury report, for what it's worth. In the other nine games where the Saints had at least 60% DVOA since 2014, Cooks always had 7-to-10 targets, so chalk it up to randomness.

Then there's the interesting case of Eric Decker with the Jets. No receiver had a higher correlation between his offense's passing DVOA and his targets (0.26), receptions (0.52), and yards (0.62) than Decker. While still not a statistically strong relationship, that's pretty damn good for what we're looking at here. Decker missed most of 2016 due to injury, but for 2014-15, he was extremely helpful to this passing offense with Ryan Fitzpatrick and Geno Smith at quarterback. He was great in the red zone and a big help in the slot to complement Brandon Marshall out wide. Our Decker sample consists of 33 games, and in the 20 games where the Jets had a positive passing DVOA, Decker caught a touchdown in 15 of them. He was a valuable player for the Jets when healthy. Now he'll try to contribute to the Titans this season.

In the spirt of Decker trying to elevate Smith and Fitzpatrick, we also have Allen Robinson and Allen Hurns trying to help Blake Bortles keep his job in Jacksonville. They have some of the highest correlations across the board as well, so perhaps it's a case where Bortles is just not capable of having a good day unless he can get Hurns or Robinson to do something significant. Robinson had at least a dozen targets in 6 of the 12 games with at least 15% pass offense DVOA. Hurns, often a slot receiver with some big YAC ability, has interesting splits with his yards, which have the second-highest correlation (0.56) with DVOA out of anyone on the list. Hurns had seven 100-yard receiving games in his team's top 14 games by DVOA. He also had 98 yards and a touchdown against the Titans last year in a garbage-time celebration. In the bottom 28 games by DVOA with Hurns on the field, he never broke 75 receiving yards. Splits happen, but that one is hard to believe.

Landry's receptions and Miami's DVOA

Every player on this list has either been a No. 1 wide receiver or a strong No. 2 for his offense. One thing I did not expect to see was any negative correlation between receptions and DVOA, but we had that with Landry (-0.14) and Watkins (-0.12). Yes, it's not a strong relationship at all, but you could suggest that as Landry's catches go up, his team's passing DVOA goes down. Compared to his peers, that's just not normal.

It is almost comical given that Landry's reception total (288) is the stat he's known best for, yet his days with big catch totals have rarely helped the Dolphins. In the 11 games with at least 50% DVOA, Landry only had more than six catches twice, including a 2014 game against Minnesota where he caught eight balls for 31 yards. In Miami's 11 worst games (-25.0% DVOA or lower), Landry had seven games with more than six catches.

Our study found 97 games where a receiver caught at least seven passes, but his offense had negative passing DVOA. Landry had a study-high 13 of those games. Demaryius Thomas was next with nine -- almost all of them coming since Peyton Manning joined the cast of The Walking Dead in late 2014. What is the significance of seven catches? That is the pace one needs to become a 100-catch receiver in a season. Yet when Landry produces at that pace, it more often than not leads to a below-average day for Miami's passing game (negative DVOA in 13 out of 21 games with Landry having seven-plus catches).

The Dolphins have been unusually good when limiting their No. 1 receiver's chances (maximum five targets), and disappointing when he is a big part of the game (10-plus targets):

Games with 0-5 Targets Games with 10+ Targets
Player Games Record Pct. Rk Avg. Off
Pass DVOA
Rk Games Record Pct. Rk Avg. Off
Pass DVOA
Rk
Julian Edelman 4 3-1 0.750 9 35.4% 9 20 19-1 0.950 1 51.1% 1
Randall Cobb 14 11-3 0.786 7 44.1% 6 12 8-4 0.667 5 38.1% 2
Antonio Brown 2 2-0 1.000 1 86.6% 1 35 23-12 0.657 6 36.9% 3
Dez Bryant 5 3-2 0.600 16 77.9% 2 9 7-2 0.778 2 34.4% 4
Steve Smith 6 3-3 0.500 20 24.8% 13 12 6-6 0.500 11 29.4% 5
DeSean Jackson 16 8-8 0.500 20 11.2% 20 4 0-4 0.000 28 29.1% 6
Pierre Garcon 16 8-8 0.500 20 20.0% 15 8 3-5 0.375 19 27.4% 7
Sammy Watkins 11 8-3 0.727 11 58.9% 4 10 4-6 0.400 17 21.6% 8
Eric Decker 6 3-3 0.500 20 -5.8% 24 13 6-7 0.462 16 21.0% 9
Brandin Cooks 6 4-2 0.667 14 45.7% 5 6 3-3 0.500 11 20.3% 10
A.J. Green 6 5-1 0.833 5 26.1% 12 15 7-7-1 0.500 11 20.3% 11
Julio Jones 2 2-0 1.000 1 71.5% 3 27 13-14 0.481 14 19.5% 12
Doug Baldwin 16 13-3 0.813 6 21.9% 14 8 3-5 0.375 19 19.4% 13
Larry Fitzgerald 9 7-2 0.778 8 36.9% 8 19 13-5-1 0.711 3 17.5% 14
Player Games Record Pct. Rk Avg. Off
Pass DVOA
Rk Games Record Pct. Rk Avg. Off
Pass DVOA
Rk
Emmanuel Sanders 7 5-2 0.714 13 41.4% 7 16 9-7 0.563 9 16.6% 15
Mike Evans 5 3-2 0.600 16 26.3% 11 19 6-13 0.316 21 14.4% 16
Odell Beckham 2 1-1 0.500 20 -20.9% 28 23 11-12 0.478 15 13.0% 17
T.Y. Hilton 7 4-3 0.571 18 -15.9% 26 19 10-9 0.526 10 11.9% 18
Golden Tate 8 5-3 0.625 15 15.9% 18 20 13-7 0.650 7 10.8% 19
Demaryius Thomas 1 1-0 1.000 1 19.1% 16 31 21-10 0.677 4 9.3% 20
Allen Robinson 6 1-5 0.167 28 12.8% 19 20 4-16 0.200 25 3.5% 21
Alshon Jeffery 6 3-3 0.500 20 -3.8% 22 14 4-10 0.286 22 2.2% 22
Jordan Matthews 15 7-8 0.467 26 -5.2% 23 13 5-8 0.385 18 0.2% 23
DeAndre Hopkins 4 3-1 0.750 9 5.8% 21 24 14-10 0.583 8 -2.2% 24
Allen Hurns 17 5-12 0.294 27 -16.6% 27 7 1-6 0.143 27 -6.0% 25
John Brown 25 18-7 0.720 12 17.7% 17 6 1-5 0.167 26 -6.9% 26
Jarvis Landry 11 10-1 0.909 4 27.0% 10 21 5-16 0.238 24 -7.0% 27
Kenny Britt 24 13-11 0.542 19 -13.0% 25 7 2-5 0.286 22 -23.6% 28

Miami went 10-1 in the games where Landry was rarely targeted, the best record for anyone with more than two such games on the list. But when Landry was targeted 10-plus times, Miami went 5-16 (fifth-worst record) and dropped in average passing DVOA from 10th to 27th (next to last) at -7.0%. Meanwhile, the Patriots with Edelman are like the opposite of Miami. The Patriots went 19-1 with Edelman heavily involved, and still had the best passing DVOA (51.1%) in those games. Now the Patriots have a lot more advantages in place to succeed than the Dolphins (and most teams), but that's a stark contrast of fortunes when these teams choose to feature their slot receivers.

Correlation between receiving yards and passing DVOA

When we got to yards, I figured Dez Bryant (0.50), DeAndre Hopkins (0.46), and Mike Evans (0.43) would do well here. They all make big plays, but more importantly, their teams follow that old-school approach of having one dominant receiver. If those guys aren't delivering, who will? I checked out Michael Irvin's prime years in Dallas (1991-1995) for this and found that his correlation between DVOA and yards was 0.38. Good, but not as high as Bryant's 0.50 since 2014. The Buccaneers might be shaking this one dominant threat trend with adding DeSean Jackson and O.J. Howard, but it has been Evans' show lately with Vincent Jackson's injuries, and he has been impressive.

Landry (0.13) had the third-lowest correlation between his yards and DVOA, but the bottom six guys all work the slot, so their usage is certainly different from their peers.

Edelman being the only player with negative correlation (-0.05) between yards and team passing DVOA was a little surprising. However, a look at the two extremes provides a good explanation again. Edelman had 141 yards against the 2014 Chargers, but that was actually one of four games (out of 39 in his sample) where the Patriots had negative passing DVOA (-3.4%). The other factor is that, when healthy, Rob Gronkowski is the biggest receiving threat on the field for the Patriots. So when Edelman only had 14 yards against the 2014 Bears and 30 yards against the 2016 Bengals, the Patriots still went over 100% DVOA since Gronkowski had 149 yards (and three scores) against the Bears and 162 yards against the Bengals. Edelman was still one of the steadiest receivers in the study, with a standard deviation of 31.6 in his yardage. That ranked second in between Pierre Garcon (28.8) and Landry (33.2).

On that note, most of these receivers had a standard deviation around 3.0 in the targets they saw on a weekly basis. A.J. Green (3.98) was the most volatile, but he's also left some games early with injury, which skewed the numbers.

Cooks (2.44) had the lowest deviation in New Orleans' always active passing game. There could be an interesting change in New England where Belichick and Brady are going to play the matchups. Some weeks will be about Cooks, but others will be about Gronkowski, or Edelman, or even the stable of receiving backs.

We mentioned that Garcon was the steadiest in yards, but he was also right up there with targets (2.86). DeSean Jackson (2.57), the only player in the study to never have 12-plus targets in a game since 2014, had the second-lowest standard deviation. The Redskins ran a pretty consistent offense with head coach Jay Gruden, offensive coordinator Sean McVay, and quarterback Kirk Cousins. With McVay and both wideouts gone, we'll see if that continues this year in Washington.

In Part II of our look at Jarvis Landry later this week, we will look at the root cause of Landry's reception value issue, and take DYAR places it has never gone before with help from this year's Sports Info Solutions game charting.

Posted by: Scott Kacsmar on 26 Jul 2017

7 comments, Last at 01 Oct 2017, 1:46pm by wiilaim11

Comments

1
by Laufy :: Wed, 07/26/2017 - 4:41pm

Good stuff - as a huge Dolphins fan and Jarvis Landry jersey owner, I tend to agree with the premise of this article, but I do have one caveat/question (that may be covered in Part 2). I'm interested to see how the depth of target affects this analysis as opposed to raw catch numbers. Based on my anecdotal perception of Ryan Tannehill, I feel like the games he struggles in often correlate to an over-reliance on quick throws to Landry at the expense of a more balanced/aggressive passing attack. It would seem to me that Landry's abilities as a receiver are being underutilized in an effort to give Tannehill a safety valve more often than not. I'm also curious how 2016 in particular compares to 2014-2015, and what (if anything) changed in the admittedly small sample size of Matt Moore's appearances last year.

2
by Scott Kacsmar :: Wed, 07/26/2017 - 9:47pm

Some of this is covered in Part II. I don't think we'll touch on Matt Moore's brief time as QB1 though.

Landry's 2016
with Tannehill - 122 DYAR, 2.3% DVOA, 108 targets, 6.6 PYD
with Moore - 52 DYAR, 16.3% DVOA, 23 targets, 5.3 PYD

Interesting. The passes got shorter with Moore, but Landry ripped off two big gains against the Jets.

3
by BigRichie :: Thu, 07/27/2017 - 11:10am

Might this be as simple as "Dolphins-or-any-team falls way behind, so opposing defense starts giving them the dink-to-Landry-in-the-Dolphins'-case passes"?

5
by ChrisLong :: Thu, 07/27/2017 - 3:04pm

I agree, this whole analysis seems to be ignoring that Landry is a slot receiver, who (probably) tend to get more passes when the team is losing, which probably means the team's offense isn't doing great. Would love to see a breakdown of DYAR/DVOA for targets by quarter, leading/behind, depth of target, etc. Also, let's keep in mind the QBs impact on these stats.

4
by johonny :: Thu, 07/27/2017 - 11:16am

With three Miami receivers on the same list makes me want to read part two. In defense of Brian Hartline-he average 14.2 ypc and was used much further down field during much of his time in Miami. Compared that to Landry 10.6 or Bess 10.7. He really was a different type of player.

6
by lightsout85 :: Fri, 07/28/2017 - 4:26pm

I'd love to hear a discussion between representatives from FO & PFF re. Landry. If the reddit-posted screenshots (from Feb 7th) are correct, Landry had PFF's 8th highest receiving grade (among WR) in 2016. And among the top 15 WRs in the screenshot, he had the 5th best Routes-Run per 1st Down.

7
by wiilaim11 :: Sun, 10/01/2017 - 1:46pm