Writers of Pro Football Prospectus 2008

Most Recent FO Features


» 2017 Offensive Personnel Analysis

It's a three-receiver league, but for the first time since 2010, the frequency of 11 personnel actually went down last year. Was it a blip, or sign of things to come?

23 Dec 2014

Any Given Sunday: Texans Over Ravens

by Andrew Healy

Very occasionally, something is so good that it is almost impossible to overrate it. You can shower it with adjectives, but you'll never quite get it right. Describing these things is frustrating because you know that someone who doesn't know about it will compare it to things they know. And they will get it wrong.

On my list of things that qualify: the Galapagos, L.A. weather, mid-1980s Wayne Gretzky, the mangosteen, The Power Broker, 2012 LeBron James, this group of doctors, a variety of things involving women, and my newest addition, J.J. Watt. Imagine a long-time fan trapped in a biosphere for the last few years without NFL access. You would try to describe how good Watt is and they would start thinking of Reggie White or some other perennial Pro Bowler. You'd show them all the stats, but still they wouldn't quite get it. Their mental image wouldn't live up to the No. 99 you see every week.

On Sunday, the real-world Watt and his defense destroyed the fourth-best team in football (according to DVOA, at least) with almost no help from the offense. Case Keenum was just as bad as the pregame line (Baltimore was favored by 6.5 in Houston) had indicated, going 20-of-42 for 185 yards and an interception against a decimated Baltimore secondary. Watt, however, helped hound Joe Flacco into a day that was even worse than Keenum's by the numbers. Flacco went 21-of-50 for 195 yards, two touchdowns, and three picks. The combined ineptitude at throwing the football was historic. Sunday's game was the first time in NFL history that both teams attempted at least 40 passes, completed less than half of them, and each threw for under 200 yards. If we look at net passing yardage instead of gross by subtracting yards lost on sacks, only one other game in NFL history fits those criteria.

While Keenum's half of the ineptitude was expected, Flacco's was not. He entered the game eighth in passing DVOA and was going up against just an average pass defense. Even with Watt's brilliance, the Texans entered the game 13th against the pass and 14th overall. The Ravens' surprising struggles on offense on Sunday were largely about Watt, but they were also about much more than that. I'll save the Watt part for last and talk about the other culprits first. Those other problems point towards the Ravens' offense being not quite as good going forward as their stats entering Sunday would suggest.

Reason No. 3 Joe Flacco was even worse than Case Keenum on Sunday: Steve Smith Fades Away

Since turning 30, Steve Smith has consistently tailed off towards the end of seasons. He has put up numbers consistent with his early-career greatness in the first half of seasons, but he has been a completely different and inferior player later in the year. This season has been no different.

Steve Smith Fades Away
Catches Per Game Yards Per Game Yards Per Target 100-yard games
First half of 2014 5.13 84.4 9.93 4
Second half of 2014 4.29 42.9 5.66 0

Smith's yards per game have fallen by about half since the season's midway point. His yards per target have fallen from 9.93, a number that would rank tenth in the league for wide receivers for the season, to 5.66, a number that would rank 96th. His first-half number ranks ahead of Julio Jones and A.J. Green. His second-half number ranks behind Riley Cooper.

On Sunday, Smith caught five passes for 49 yards, averaging just 4.1 yards on 12 targets. It was Smith's fifth time in eight games averaging less than 5 yards per target. Given that Smith's late-season swoons are a recurring pattern, the odds are that Sunday's Smith is the version we'll see the rest of this season. The Steve Smith, Sr., who was a big part of the Ravens' best offensive performances of the season (against Carolina, Tampa Bay, and Miami) may come out, but it is not particularly likely.

Reason No. 2 Joe Flacco was even worse than Case Keenum on Sunday: Flacco's 2014 uptick is pretty random.

From 2008-2012, Joe Flacco was about as consistently average as a quarterback can be. He ranked between 15th and 22nd every year. Then that apparently unshakable mediocrity showed a lot more variation. First, Flacco had that remarkable playoff run in 2012. Then, Flacco was terrible last year, ranking 35th in DVOA, before rebounding this year with his first top-ten season.

While some things (pass protection) have arguably gotten better for Flacco this year, others (his weapons) have arguably gotten worse from his pre-2013 seasons. Rather than changes in his supporting cast or his skill level, simple randomness can account for Flacco's uptick. When we try to project Flacco from here, the data mostly suggests that we're going to see the same Flacco we have seen for seven years rather than the version who looked different over a smaller sample this season.

Projecting Flacco's DVOA for next season using a sample of quarterbacks with at least five years of previous experience with a regression gives these results over 10,000 simulations:

Projecting Joe Flacco for 2015
Mean DVOA -1.18% Chance Flacco Ranks in Top 10 17.0%
Mean Ranking 19.9 Chance Flacco Ranks between 11 and 20 33.7%
Chance DVOA is Negative 58.1% Chance Flacco Ranks Between 21 and 30 33.1%
Chance DVOA is Positive 41.9% Chance Flacco Ranks 30th or Worse 16.2%

Flacco, relentlessly average for five of his first seven years, projects to be average next year, too, and just as likely as any average quarterback to have a stinker against a good-but-not-great defense such as Houston's.

Reason No. 1 Joe Flacco was even worse than Case Keenum on Sunday: J.J. Watt

Sunday was Watt's third game this year against a top-ten offense, and his first Watt-goes-off-level dominant performance. He had eight tackles for just the fourth time in his career, three tackles for a loss, one sack, and four quarterback hits. Watt got credit for none of those on Joe Flacco's second interception, even though he caused the poor pass.

Even though Eugene Monroe took Watt down, Watt still got close enough to Flacco's feet that the quarterback did not step into the throw, leading to the interception that set up Houston's first touchdown. Even Watt's otherworldly numbers somehow don't do justice to the way he affects opposing offenses. It is impossible to quantify how much a quarterback's mental state changes when Watt enters unstoppable mode. His impact was the driving force behind the Texans' huge day on defense, a day that overrode the Texans' offensive impotence.

BAL -70.3% -28.9% 15.4% -26.0%
HOU -29.6% -89.1% -5.8% 53.6%

BAL -82.1% -30.7% 15.4% -36.0%
HOU -39.3% -79.6% -5.8% 34.5%

But Still, Let's All Stay Calm

Even though Watt is the league's best defensive player by an unusually wide margin, he's still not the MVP, and it's not close. For a bunch of reasons, I think his candidacy is substantially weaker than it was in 2012. It's more than his defensive stats being a little less incredible this season than his 2012 numbers, or the Texans winning at least three fewer games this year, or that he was the best defensive lineman in the league last year and still his team won just two games and finished 30th in DVOA.

In addition to this obvious stuff, two other reasons to oppose Watt's candidacy have received less attention.

  • Even after the game against the Ravens, the Texans' defense as a whole group is still only good, not great.

For all his brilliance, Watt has only lifted the Texans' defense to slightly above average. Think of the offensive floor for a team with a great quarterback. Over 20 combined seasons since 2005, offenses led by Rodgers/Manning/Brady have an average DVOA ranking of 3.4 and have ranked no lower than ninth. Brady and Manning over that time never ranked lower than seventh, and that includes a Brady season with Reche Caldwell and Jabar Gaffney playing receiver. The Texans this season have a -6.3% defensive DVOA, which puts them closer to 19th-ranked Kansas City than to top-ranked Detroit.

  • Unlike 2012, Watt has wreaked much of his wrath this year against the worst offenses.

The Texans have played seven of their 15 games against offenses ranked 25th or worse by DVOA. Great players dominate bad offenses just like great teams dominate bad ones, and we definitely don't want to penalize Watt for destroying the Jaguars, Titans, and Bills (against whom Watt had an almost unbelievable nine quarterback hits to go with an 80-yard interception return). Still, we need to correct Watt's pass-rushing numbers a bit to account for his schedule. This year, Watt has averaged 1.3 sacks and 4.0 quarterback hits in his seven games against the bottom quarter of the league, compared to 1.1 sacks and 2.4 quarterback hits against the rest of the league. All those numbers are great, but we'd project Watt to have about one fewer sack and about seven fewer quarterback hits if he had played a league average schedule.

In 2012, Watt was almost identically effective in sacks (1.3 per game) and tackles (4.3 per game) against the worst eight offenses and the rest of the league. So I just keep coming back to this being the wrong year to give Watt the MVP. Putting aside his plays on offense, 2012 Watt was a little greater than 2014 Watt. The 2012 Texans' defense posted a better DVOA of -14.2% and ranked three spots higher. And the writers even were in the mood to do something crazy, as they gave the most valuable player award to one of the least valuable positions on the field. It would have been the wrong call to give Watt the MVP in 2012, but it would have been a better call than giving it to Adrian Peterson. I don't want to see Watt win the award this year when he still plays the wrong position, when he has been a little-less-great against a weaker schedule, and when the Texans' defense is not quite as good. It may be impossible to overrate how good J.J. Watt is, but it still is possible to overrate his value.

Posted by: Andrew Healy on 23 Dec 2014

6 comments, Last at 26 Dec 2014, 3:06pm by LionInAZ


by PaddyPat :: Tue, 12/23/2014 - 3:54pm

The trouble with the Watt MVP analysis here is that comparing Watt to himself is really the wrong approach. You need to try to compare him to the other potential candidates this year. I mean, there were several years when Manning won MVP where he was a mere shadow of what he had been in other years, but he was still a strong candidate for best performer in the league THAT year.

From your general analysis, it would seem that there's really only a poor case to be made that any position should win MVP outside of signal-caller, and I suppose that's fair in this modern passing league. After all, receivers rarely elevate their quarterbacks; running backs are largely interchangeable, and even fantastic offensive and defensive lines can rarely compensate for a poor passing attack--ask the Bills, or even the Jets.

With all of that taken into account though, which signal caller truly defined his team's success this season in a manner that impacted the league and was demonstrably the Most Valuable?

Peyton Manning has faded somewhat this year, and is on a team that has increasingly relied on the run to complement a great defense.

Tom Brady has (distressingly frequently) done only just enough to help out a defense that is winning games for him. He has had some great moments this season (as has Manning) but it has not been consistent.

Philip Rivers likewise has had some fantastic stretches and some weak ones, an argument that could also be made for Tony Romo.

Aaron Rodgers is probably the best candidate. He has been efficient, avoided mistakes, made big plays, and performed well against good opposition as well as weak. The losses have been rather striking though. Stinkers at Buffalo, New Orleans, Seattle and Detroit, all suggest a player who has been a little less than Transcendent. If he roasts Detroit this coming weekend, he has a great case. If he struggles again, however, I think he falls rather short.

All of which combines to suggest that if the MVP is not Watt, then who is it this season? It may not be Watt's best year, but he has been consistently phenomenal, week in and week out, elevating a defense that is middling around him and giving his team a shot at the playoffs entering week 17, and come on, what else does that team have going for them? Watt's impact cannot be overstated. In a season where the passers have been good but mortal, Watt has been an immortal player, outperforming the rational and reasonable expectations of his position.

by Andrew Healy :: Wed, 12/24/2014 - 2:23pm

I think Rodgers would be a deserving MVP. He's now second in DVOA. And his line is only 15th in adjusted sack rate. His team is 4th in DVOA, compared to Houston being 20th. But you're certainly right that there isn't much around Watt. Still, even in their worst years, Brady and Manning had offenses better than Watt's defense this year. Is the Houston defense supporting cast worse than the Patriots' 2006 offensive supporting cast?

And Rodgers has had some bad games, true, but even Watt came pretty close to a Gholston against the Raiders in his defensive stats (0 tackles, 0 sacks, 1 QB hit).

by DisplacedPackerFan :: Tue, 12/23/2014 - 4:25pm

See, I think you missed it with the Reggie White comment. When I try to describe how good Reggie was, because like you said there are some things that are just hard to make people get, the only way I can get them to is to say, "Watch J.J. Watt play and then make him just a touch better." I watched LT and Bruce Smith destroy teams but didn't think either was on Reggie's level. I love J.J. Watt because he is the only player I've seen that compares to what Reggie did.

I agree with most everything else you said. Watt is just crazy good. I think you undersell White, by a lot, though.

I also tend to think of MVP as MOP (Most Outstanding Player) and like PaddyPat said so well, if you don't give it to Watt, who do you give it to? I love Aaron Rodgers but some of his lows have been low. Of course they were no worse than some of the lows that past MVP winners have had, so I wouldn't say he was a bad choice, and I think he is still the best choice among the QBs. As stated, the Detroit game will have a big impact.

by jonnyblazin :: Tue, 12/23/2014 - 5:15pm

I strongly disagree with the statement that Flacco's 2014 uptick is random (however, Sunday's game will surely drag his DVOA down quite a bit). Flacco is not a QB who makes calls at the line or audibles, so his success is tied with the quality of the offensive coordinator, as well as having reliable offensive weapons who he can trust to make a play on the ball in 1-on-1 situations. Pretty much every season indicates this:

Looking at the DVOA:

2008: -3%
2009: 8.4%
2010: 9.4%

The first 3 seasons Flacco improved under Cam Cameron, and had his best season when the Ravens brought in Boldin to complement D. Mason and T. Heap, who were aging and not very effective by 2010.

2011: 0%
2012: -1.3%

The Ravens jettison Mason and Heap, and Flacco eventually develops a comfort level with Boldin and Pitta to coincide with the Super Bowl run. Keep in mind Flacco's 2012 DVOA would look much different with the 4 playoff games (11 TD to 0 INTs), I'm still not sure why that data is not included in FO stats. Jim Caldwell replaces Cam Cameron late in the year, and runs Cameron's system but varies the playcalling to great effectiveness.

2013: -18.1%

The Ravens jettison Boldin, and Pitta misses the season, and Jim Caldwell turns back into a pumpkin. Turns out there's a reason Caldwell has never run an offense before, teams were calling out the Ravens plays by the end of the season.

2014: 12.2%

The Ravens bring in a good offensive coordinator in Kubiak, who runs a WCO offense that is generally more favorably regarding by DVOA. The Ravens sign Steve Smith and Owen Daniels to solidify receiving corps.

That seems anything but random to me. There's pretty strong evidence that Flacco can produce at an above-average level, but he needs a good OC and veteran, reliable receivers to succeed. Cam is a good developer of younger QB's, but his game planning and playcalling are limited so the upside to his offenses is not very high. Caldwell should never run an offense again, and Kubiak is much better than Cam and Caldwell.

Flacco is pretty good at identifying 1-on-1 matchups down the field, but when throwing to guys like Torrey Smith there are problems because Torrey easily gets outmuscled for the ball or falls down. Flacco worked well with Boldin, and to a lesser extent Steve Smith, because they will consistently fight for the ball, even though they are not elite WR's.

by SFC B :: Wed, 12/24/2014 - 7:23am

Texans fan here so keep that in mind with my comment here, but I'm of the belief that if Watt's candidacy for MVP this year can't be championed then we might as well do away with the MVP honorific and go with Defensive Player, Offensive Player, and QB of the year.

by LionInAZ :: Fri, 12/26/2014 - 3:06pm

I'd go further and have awards for each position, so we can be spared all the arguments about which position is the most valuable.