Gary Kubiak, Empty Backfields, And You

It's third-and-1 on the Cincinnati 18, with 51 seconds left in the first quarter. The Texans are driving, threatening for their second score of the game.

Arian Foster goes into motion towards the outside hash, leaving the backfield empty.

If you are anything like me, this idea drives you up the wall. I don't care that the Texans were able to get a first down on this particular play (a Matt Schaub dive, which is a play Kubiak calls once in an Albert Haynesworth-sized moon). Removing the threat of your Pro Bowl running back taking the ball behind an offensive line with the likes of Duane Brown and Chris Myers on it just feels incredibly counter-intuitive. Passing it through my video game filter, it's kind of like playing an RPG and making all of your characters be bards -- sure, it might work, but why be so cute about it? Schaub's ability to stand and deliver under pressure is oft-criticized (if not entirely fair), and with the Texans possessing fewer than three receivers who can beat press-man coverage, it just seems like Bad Idea Jeans.

So once again I fired up the handy-dandy Football Outsiders' play-by-play spreadsheet and looked at the entirety of the 2012 season. Since I was only trying to prove things about the Texans, I didn't pull out the league for comparison like I did in last week's post. This is all Texans-related data.

• Of the plays Football Outsiders charted for Houston, the Texans had a running back out wide on a pass play 37 times. That means we are working with a sample of about 6.3 percent of all of their passing downs on the season. In other words: we are going to be running into small sample sizes again. There were another four or five run plays with a running back out wide, but I'm going to ignore them like Kubiak ignored the idea of putting Ben Tate and Foster in the same backfield after trying it one time in Week 1.
• Of those 37 plays, only 13 of them were "successful" plays by FO definitions. (On first down, 40 percent of needed yards for a first down. On second down, 60 percent of needed yards for a first down. On third or fourth down, 100 percent of needed yards for a first down.) That's a success rate of 35 percent. They averaged 4.9 yards per play on these throws.
• Breaking it down a little further, the Texans were 4-of-18 converting first downs or touchdowns on third down passing out of this package. At this point we're into major small sample size territory, but just to state them for the record: the Texans averaged 5.8 yards per pass on third down with a running back out wide. They also averaged 1.6 yards per pass on second down, and 8.4 yards per pass on first down under the same conditions.
• Now, for the part that really rankled me: the Texans did do poorly on third-and-short (I'm going with third-and-6 or fewer yards to go) when they put a running back out wide. However, it's hard to get too fussed up about it because it only happened six times over the course of the entire season. They were one-of-six converting those opportunities.The surprising part was that when I went back and looked at this in 2011, seeking a bigger sample size, the Texans actually got a first down or touchdown on seven-of-12 third-and-short throws that season. In fact, while it happened on second down, you may remember this play that occurred with Foster going out wide:

It was a pretty good play. Overall, the Texans were successful on 19-of-40 pass plays with a running back out wide in 2011.

• How about the idea that the Texans did this because they weren't very good running the ball on third down? Of the Texans' 68 third-down runs last season, only 26 of them were first downs or touchdowns. That's a success rate of 38 percent. Not much better than they did by putting someone out wide. However, if you remember what we talked about last week, you'll see that a lot of those came on third-and-long for some asinine reason. If we limit ourselves to the same third-and-6 or fewer threshold we used above, the Texans succeeded on 25 of their 45 carries. That's a much more respectable looking 55 percent. It's also a higher success rate than Foster's 47 percent season average, meaning that the Texans actually did slightly better running on third down when they actually had a prayer in hell of converting. (For the record, the overall success rate on third-and-short -- runs and passes -- was 51 percent.)

So what do I glean from the numbers? That as much as the idea of this play-call is aesthetically unpleasing to me, it's really not that big of a deal. Sure, I'd rather that it happened on one of every 40 passes instead of one of every 20 passes, but factoring in game theory and the idea that you want to make your opponent prepare for something, it's probably not something worth making a huge fuss about. When you bring in the facts that (a) they were actually pretty successful with it in 2011 and (b) we're working with incredibly small sample sizes, going one-of-six is not a totally crazy result that spells out doom and gloom for any who would dare run the play.

That said, it's still a stupid play that I hate and want to die, and I will still harp on it every time it doesn't work because WHY ARE YOU LETTING THE DEFENSE KNOW AHEAD OF TIME THAT THEY DON'T HAVE TO WORRY ABOUT THE SECOND-BEST SKILL POSITION PLAYER YOU HAVE VERY MUCH AT ALL?

But, from a purely analytical perspective, it's a functional -- if not phenomenal -- package in Houston's offense. The bashing we gave it in Hair Of The Dog all season was probably unwarranted.

Just don't ask me to be logical about it. That's all I ask.