The return of 2017

A reason that I was so down on the Texans before this season, and then only kinda dragged against my will into thinking they’d win the division, was because a core belief of mine is that Bill O’Brien wasn’t a very good head coach last year. He showed the ability to make adjustments at opportune times — I’ve used the Watson read-option stuff in the Jags Week 17 game as my go-to for this all offseason — but he went right back into the long-developing garbage chute for the Colts playoff game.

So it’s worth pointing out that the reason the Texans have looked so good the last few weeks isn’t just the growth of Deshaun Watson — it’s because the head coach made an adjustment that has worked on every level, and righted one of the wrongs of the past. For that, O’Brien is getting a lot of credit from me.

2017 Deshaun Watson was never actually stopped

Deshaun Watson finished seventh in DVOA in his rookie season, 2017. He did this not in the offense that the Texans ran in 2018, but in the offense they molded to a lot of his college concepts. While he was not perfect, he showed a lot of promise. I will hand this over to Charles McDonald because his piece on it for Football Outsiders is the first one I thought of:

We’re only in the early stages of Deshaun Watson’s career, but it should be noted that he has improved every time he has stepped on the field through the first three weeks of the season. Some people might say that his day only came because the Texans were playing a poor pass defense, but good players routinely take advantage of inferior matchups on the football field. The Texans have to be very, very pleased with the early progress their young quarterback has made so far.

Watson only got better from there. In Weeks 4-8, before Watson tore his ACL in practice after the Seahawks game, the Texans had a positive pass offense DVOA in every single week. It was as high as 84.7% against the Chiefs in Week 5, and as low 21.4% against the Browns in Week 6. The Texans weren’t always able to mesh it with a good running game because their line was so bad at this point, but Watson’s ability to play off RPO and read-option games gave them a lot of extra space.

If you dropped that play into last week’s Kansas City tape, it would have been essentially in place. That’s what a lot of Watson’s best 2017 stuff was about, because he wasn’t as good then as he is now. There were a lot of schemed quick throws that were empirically successful.

Bill O’Brien removed the easy plays from the playbook in 2018

I wrote about the beginnings of the 2018 season and the sense of what had changed for Houston for The Athletic.

The short of it is that O’Brien preferred longer route combinations to the offense’s detriment in 2018. In researching for Football Outsiders Almanac 2018, I could only find a handful of actual, legitimate option plays that Houston ran. When they did run them, they worked. The one game in which they used more of them than in any other game I charted was when they went to Indianapolis, scored 38 points around a bunch of interesting Keke Coutee stuff, and won.

And remember, this wasn’t with O’Brien having seen a lot of tape about Watson getting owned while teams sniffed out the read-option dressing or anything. They were dominant offensively using that in 2017. He chose to keep that out of the offense as a base look.

The start of 2019 was more of the same

My theory is that a lot of the reason behind trading for Kenny Stills and Laremy Tunsil was about running the O’Brien 2018 offense.

Houston’s adjusted sack rate actually got worse through four weeks despite empirically better line play. Houston’s quarterback hit rate declined, but only by a tiny amount.

While O’Brien did introduce more RPO elements into his offense than he did the year before, the focus of the team continued to rely on Watson to create out of structure to make offense happen. Too many of Houston’s big plays were “Watson buys time for receivers to get open.” It remained a big problem even in games like Week 3 against the Chargers, where Watson wasn’t pressured all that much. There were still plays where he had to buy a lot of time.

Watson is an aggressive quarterback, O’Brien is an aggressive head coach. It always made sense for them to aim downfield, but they never seemed to have a plan B for that in 2018 or the beginning of 2019. If they missed the deep throws, we would sometimes be watching a team with a ridiculous amount of talent struggle to hit 20 points.

What happened to the Falcons and (especially) the Chiefs

While Houston’s turn to 12-personnel in Week 5 opened up the play-action game, Week 6 was a full-on embrace of 2017’s principles, and with better talent and smarter ancillary designs than we saw in 2017. They didn’t hit a single deep ball and they owned the Kansas City defense so bad that they never punted. The Texans have punted once in the last two weeks.

Here’s them utilizing Darren Fells as a drag player in a read-option play:

I understand that it’s a very NFL fan thing to look at option plays and be like “well, that won’t work in the NFL for long.” Well, the NFL isn’t that much different than college football. The main difference isn’t the strategy, it’s the speed of the players. I know we can get way deep into the weeds about how this play had Akins as a wheel route guy on this play, and I could talk over everyone’s heads for about two paragraphs. Let’s just try to keep things simple:

Option plays and play-action plays share a common principle: Making a defender guess. If the defender corrects, and is one step behind, that’s a big hole in NFL terms. One step behind is a crease Carlos Hyde can run through. One step behind is a mostly-open receiver. If you build an offense around this stuff, and a defense can’t figure it out, it’s free. That throw to Fells was 10 free yards. And then you’re in their head:

Carlos Hyde starts to benefit. Tight ends blocking defensive ends start to benefit. Every player gets that little extra bit of help, and all of the sudden you’re running an offense that is overwhelmingly successful. That’s why I was such a stick in the mud about O’Brien abandoning this stuff in 2018. It never made any sense when your quarterback is one of the five best athletic marvels in the league — defenses actually have to play him honestly. Nobody has to play Jared Goff honestly as a run threat. They barely play Carson Wentz as a run threat, and he can actually run. The stuff Lamar Jackson is doing in Baltimore is terrifying to defenses.

Then you involve Keke Coutee, and now the defense is literally trying to read four different players out of the backfield, all of which are either blocking, eligible receivers, or running threats. (Well, Fells probably isn’t a running threat, and Watson probably isn’t blocking, but you get the picture.) You’re putting a lot of mental pressure on the defense to read things correctly, with lots of cheap yards in front of you if they can’t.

It never made any sense for this to not be the way forward for this offense. It was loaded with playmakers at the skill positions, and this offseason brought a fresh batch of tackle breakers (Duke Johnson) and deep threats (Will Fuller’s return to health, Stills) that would make play-action dangerous. More to the point, Watson gets hit less running this offense than he ever did in 2018 — involving him in run-action is less dangerous for him long-term than what we saw early this season.

It’s also worth noting that this doesn’t have to express itself solely as read-option/RPO principles. A lot of what is important is giving Watson early quick reads that let the offense do the work they’re supposed to do. There were more quick curls in that Falcons game than I can remember seeing O’Brien run in a long time.

Over the last two weeks, Watson leads all quarterbacks in DYAR and has a 56.6% DVOA. He understands the field better than he did in 2017. He’s less turnover-prone, and, as O’Brien noted in his Monday presser, making smarter reads.

The Texans didn’t even play that well against the Kansas City defense. They left meat on the bone on a few running plays. Fuller got open deep three times and caught none of them on his person. Hopkins alligator-armed a ball on the goal line. And they still scored 31 points and marched that defense down the field until it was depleted. Then they stuffed fourth-and-3 right down Kansas City’s craw:


Listen, it’s too soon to say whether these changes will stick as a base game plan. Stills could come back and we could be right back in 11-personnel blitz-to-the-flat ball. The Cowboys had two great games of play-action to lead off the 2019 season and Jason Garrett dragged them right back to the gutter. The Rams have never completely figured out how to adjust off Goff’s play-action being taken away. I’m not saying that as an insult to O’Brien — I just think it’s important to note that head coaches change game plans on a weekly basis and they aren’t flawless reasoners. There’s also the fact that everything is figureoutable in the NFL to some extent, and somebody is going to figure out how to defeat this stuff against the Texans someday, even though it hasn’t happened yet.

But we now know two things, two very good things, that we didn’t know in Week 1.

— Bill O’Brien is aware that basing a lot of the offense off options will kick ass for him.
— Bill O’Brien saw enough on tape to understand that trading for a left tackle alone wasn’t going to fix where Watson was at in dealing with this offense circa 2018.

Those are both objectively good things and reasons why O’Brien deserves a lot of credit for this change. It takes a lot to make a midseason change like that, and I can only hope it isn’t an underdog tactic and instead becomes a base part of the game plan.


I’m happily writing this article free of charge — this is a labor of love as I am between Texans gigs. This is presented to you ad-free and without any hassle. If you enjoy my work and want to encourage me to produce more, please feel free to leave me a PayPal tip.

One Reply to “The return of 2017”

Leave a Reply

Your email address will not be published. Required fields are marked *