How Successful Was The Cowboys Offensive Line Against The Splash Play?

The fewer splash plays an offensive lineman allows, the better. Today we'll look at how many splash plays the Cowboys' linemen gave up and compare that against the NFL average in 2012. But before we start simply adding up the number of splash plays, it's important to understand that pressure is not distributed evenly along the line. Tackles obviously give up more pressure than the interior linemen. But even if you had the average sacks given up by a left tackle, you wouldn't necessarily be able to compare one left tackle to the other because one guy may have played 300 pass blocking snaps, while the other guy played 600 snaps.

Example: Nate Livings gave up 14 QB pressures according to Pro Football Focus (PFF), while the Patriots' Logan Mankins for example only gave up 11 pressures. Does that make Mankins the more effective pass protector in 2012? Some people would argue that that's the case, heck, some people would even argue that point without looking at any numbers. But Mankins only played on 426 pass blocking snaps, while Livings was on the field for 700. The average NFL lineman played on about 600 pass blocking snaps in 2012, so if we adjust for that, Mankins' "pressures-allowed-per-600-snaps" climbs to 15.5 adjusted pressures , while Livings' total declines to 12.0 adjusted pressures. Based on this specific metric, adjusted pressures, Nate Livings was better at preventing QB pressures than Logan Mankins in 2012.

I performed a similar 'normalization' to 600 pass blocking snaps for all NFL O-linemen who played at least 25% of their teams' total snaps. This is what the average NFL line gave up in total pressure in 2012, based on 600 pass blocking snaps:

The pressure is bigger from the outside, no surprise there. The days when O-lines and QBs were more susceptible to pressure from the left- or blind side are long gone. For one thing, more left-handed QBs in the league balance those numbers a little. But also teams now tend to put their best athletes at LT to counter the pass rush, and the best pass rushers are lining up where they can create the best mismatch, and that may not be only on the left side anymore.

Now that we've established the baseline, or 'average' for these splash plays, let's have a look at how the Cowboys linemen compare. To compare the Cowboys linemen with the NFL average per 600 snaps, I used percentages in the table below to show how much better or worse they are versus the NFL average. 100% means a player is exactly on average, 50% means the player has allowed only half as many splash plays as the NFL average, 200% means he's given up twice as many. Obviously, the lower the percentage, the better.

To improve the readability of the table, green cells denote better than average performance, yellow is up to 25% worse than the NFL average while red is more than 25% worse than the NFL average at the position.

The first thing that pops out here is the column on the far right which sums up the performance of all five linemen, and which shows that in terms of total splash plays allowed, the Cowboys O-line was right around the NFL average. As a unit, they were very close to the average in terms of sacks allowed, better than average on QB hits and worse than average on QB pressures allowed.

If this feels intuitively wrong, keep in mind that these numbers are adjusted for the number of pass-blocking snaps the Cowboys O-line played, and the Cowboys had a heck of a lot of those, ranking third in the league in pass attempts. We can use more traditional stats to illustrate the point: The Cowboys allowed 36 sacks last year, tied for 15th in the league. But their sack percentage [SACKS / (PASS ATTEMPTS + SACKS)] of 5.2% ranks them number 10 in the league. That is not a bad number.

The other reason why this may feel intuitively wrong has a lot to do with something called the "Availability Heuristic" or "Availability Bias", which is based on the notion that, "if you can think of it, it must be important." Here's how that works: Try to visualize what you remember about Doug Free's performance last year. If you're like me and probably 98% of all football fans, chances are the only thing you can visualize about Doug Free is an image of how he's helping up Tony Romo after somebody ran by Free for a sack on Romo.

Per Stats Inc., Free allowed seven sacks last year; PFF says Free allowed six sacks. That's six or seven plays out of 640 pass blocking snaps that Free played last year, or about 1%. Effectively, for most football fans, the mental image of Doug Free as a pass protector is based on 1% of his total body of work. And yes, when people talk about the "eye-test", the underlying mechanism is nothing more than the availability heuristic.

Don't get me wrong, I'm not saying Free had a good year, not by a long shot, but the numbers above make a strong case that the Cowboys' O-line - as measured by the adjusted number of splash plays - was an average unit. Nothing more, nothing less.

Back to the numbers: The second thing that pops out about the table above is that the numbers for the individual players are all over the place. Smith looks good on sacks and hits but was the worst Cowboys lineman in the relative number of QB hits given up. Livings looks like the exact opposite, bad on sacks and hits but good on pressures. Same thing for the other linemen: each one has a different combination of above average and below average performance. And it is perhaps this lack of consistency that was the biggest weakness of the 2012 O-line. If you have one weak link in the line, you could maybe try to scheme around that weakness, but when you spring holes all over the place, then it gets tricky. You may be able to plug a hole, but you can't plug a sieve.

The next thing to look at are penalties. As penalties are not limited to the passing game, I've normalized the number of penalties to 1,000 snaps, which is about the average number of snaps a lineman plays over a 16-game season. Here's how the Cowboys look against the average:

The key issue here are both tackles. Free got flagged 15 times, Smith 11 times, which makes Doug Free the most penalized offensive lineman of 2012, with Smith ranking 5th. That's not a good place to be in. This may be hard to believe for Cowboys fans, but of the 80 tackles that played more than 25% of their teams' total snaps, exactly half were flagged only five times or less. It's not impossible to play tackle without incurring so many penalties.

Overall though, what these numbers do not show is the quality of splash plays allowed. Smith's TD-saving horse collar penalty for example was a good penalty to get. And it's hard to say from these numbers whether a sack was a coverage sack or whether it was a sack where the lineman was beaten like a rented mule. And the numbers also don't account for the amount of times that Romo spun his way out of a sack, just as they don't show the number of times that Romo had seven or more seconds to complete a pass.

What we're left with is a unit that performed roughly at an NFL average level against the splash play - at least according to these metrics - but whose biggest weakness may have been the inconsistency across all five positions. Ultimately, the Cowboys' O-line didn't have any winners against the splash play. We'll find out over the coming months what the Cowboys plan to do about this.