This is part one of Poseur’s two-part series for offensive line statistics.
One of the hardest things to evaluate in football is line play, just because there is a complete dearth of analytical tools at our disposal. We are pretty much at the mercy of common wisdom, which becomes a sort of group think. Team X has a good line because one expert said so, who is usually parroting another expert who is parroting another. And our statistical toolbox is virtually empty for evaluating lines.
I’m trying to rectify that. Now, I’m not claiming my o-line states are flawless. No statistic is, but it at least gives us something to start a debate with. It gives us a context to objectively evaluate the lines. Before I unveil the system, let’s admit the difficulties and flaws of creating an objective statistic for offensive line play.
First, an offensive line works as a unit, not as a bunch of individuals. At no position is the cliché “only as good as the weakest link” as true as it is with the offensive line play. So I don’t even attempt to evaluate individual linemen. If you want to know how good Ben Grubbs is, I really can’t help you. The system here is to evaluate each unit as a group.
Secondly, everything an offensive line is dependent on someone else. The line has essentially two jobs: prevent sacks and open up holes for the running game. So by counting the sacks or the average yards per carry is not just the responsibility of the line, it is determined by the mobility of the quarterback or the skill of a running back. This is an unavoidable fact of ANY football stat. Every statistic is influenced by more than just the player being evaluated. Receptions are a function not of a receiver’s skill but of the quarterback’s skill and the coach’s decision to throw the ball. Hell, every running stat has been dependent on the offensive line and we don’t throw those numbers out, so turnabout seems like fair play. Isolation of a skill in football statistics is frankly an impossible task.
Finally, opportunity matters. The number of sacks allowed for a pass-happy team like Kentucky is not the same as the run-oriented offense of Arkansas. This is actually a fairly easy problem to rectify, as we will look solely at rate stats. It’s not raw sack totals or raw rushing yards, but sacks per attempts and yards per attempt. This puts teams on an equal footing regardless of their offensive philosophy.
So, onto the system itself.
Like I mentioned, the offensive line has essentially two jobs: preventing sacks and opening up holes for the running game. Which means we can evaluate the line in two separate facets of the game: the running game and the passing game.
It’s actually pretty easy. Take the total number of allowed sacks compared to the total number of pass attempts. OK, not exactly because there is some slight modifications needed. In college, sacks count as rushing attempts, not pass attempts, so the sacks have to be added to the total number of pass attempts. Also, to put the number on a more readable scale, we multiply the result by 100. It leaves us with this formula:
100*Sacks/(Pass Attempts + Sacks)
Taking the average SEC team, we plug in these numbers: 24 sacks and 359 attempts (actually, 24.167 and 358.583). The average team allows 6.739 sacks per 100 pass attempts.
Well, we can’t use simple yards per attempt because of the problem with sacks. Sacks count as rushes, not as pass attempts and we need to avoid penalizing a team twice for allowing a sack. The modified yards per rush formula we’ll evaluate offensive lines purely on rushing with is this:
(Rush Yards – Sack Yards)/(Rush Attempts – Sacks)
Once again, taking the average SEC team, let’s plug in some numbers. The average team attempts 435 rushes on the season, gains 1815 yards, and allows 24 sacks for 165 yards (actual figures: 435.333 rushes, 1815.083 yard, 24.167 sacks, and 165.083 yards). The average team averages 3.938 yards per attempt.
PUTTING IT TOGETHER
We’ve now encountered another problem. The numbers are on completely different scales. Take a look at the average team: 6.739 sacks and 3.938 yards/rush. If we work on the assumption that pass blocking is equally important as rush blocking, it’s impossible to simply add those numbers together. The ratio is too bizarre.
What we need to do is get these two variables on the same scale. And here we turn to one of our friends from your college statistics course: standard deviation.
We can take figure out how many standard deviations a team’s performance has been from the mean in both run blocking and pass blocking. This puts both variables on the same scale and also makes the numbers completely dependant on how the rest of the league has performed. We’ll call each computation of the number of standard deviations the PSCORE and the RSCORE.
Then, it’s a simple matter of adding the two numbers together to get an OLINE score. To keep the total number from being a long decimal, I multiply the result by 100 just to make it more readable.
This also means that 0 is a perfectly average offensive line. Any number in the negative is a below average team, and any number in the positive range is above average. Since this is running a bit long, we’ll put this into practice and apply these formulas to last year’s SEC teams in the next installment. There are some surprises on who ranks as a quality and poor offensive line. But without spoiling the suspense, Arkansas’ line was not just awesome, it was fucking awesome. It was every bit as good as we thought it was.
So at least the numbers passed the early smell test.