Averages are among the most misunderstood stats in sports, yet we rely on yards per carry, yards per pass attempt and yards per catch as indicators of performance. In the Baltimore Ravens’ case, they averaged 4.5 yards per run and 7.2 yards per pass. Presumably, then, if the Ravens ran the ball, the most likely result was a 4.5 yard gain, right? Hardly. Likewise, the most likely result on any given dropback was most certainly not 7.2 yards.
Discuss on the BSL forums here.
After compiling the data, I learned that averages tell a very limited part of the story. Take a look at these histograms, one for the ground game and one for the passing game, and see the frequency of a given result on a play.
To build these graphs, I started by compiling the result of every single offensive play from scrimmage this year. I didn’t include any penalties aside from defensive pass interference, because the Ravens relied on it as part of their offense, and intentional grounding, because it is akin to a sack. I split the data into two charts, one for running, one for passing. Then I counted the frequency of a given result (for instance, the Ravens ran for 11 yards eight times, 12 yards three times, etc.). Then, I had them turned into a histogram to illustrate the data.
Let’s start with passing plays.
Data compiled by Shawn Brubaker. Chart made by Dan Bryden. Please credit accordingly if reusing.
As you can see, by far the most common result was a gain of zero, either an incompletion or a completed pass for zero yards.
The most common result for a completion, though, was seven yards, with passes between five and 15 yards being very common results on the whole. Surprisingly, there were fewer total negative yardage plays through the air, in part because Joe Flacco was so rarely sacked.
That said, the frequency of zero-yardage plays challenges the conventional wisdom that passing is usually the better option. The opportunity for a gain above 5 yards is much higher, but the chance of a turnover, no gain or loss of yards is also much higher.
Now on to rushing.
Data compiled by Shawn Brubaker. Chart made by Dan Bryden. Please credit accordingly if reusing.
Here the data was more unexpected. The most common results, in order: one-yard gain, two-yard gain, three-yard gain, no gain. In all, a VAST majority of plays went for less than what the Ravens actually averaged.
Even against the New Orleans Saints, a game in which the Ravens ran the ball very effectively, close to two-thirds of the Ravens’ runs went for three yards or fewer.
It should surprise no one that big plays were critical in boosting the Ravens’ average. Just how critical, though, was shocking. The idea of consistently gaining four or five yards consistently per carry isn’t a reality in Baltimore. Whether it is a realistic expectation for any other team is questionable as well.
To be honest, my hypothesis in doing this experiment was that passing would produce more big plays (which proved true), but running would be more consistently productive. That was not true. On the contrary, passing the ball was a far more consistent way to gain yardage. Really, the only downside to passing is the increased chance of a turnover (which is admittedly a very big downside).
That said, though running the ball is statistically less effective than passing it, the ground game remains crucial. There is no doubt just from glancing over the raw data that the Ravens pass the ball much more effectively when they’ve established a ground game. That’s something I will analyze more closely in a future article.
There is much more to be gleaned from this information than what I have done so far. I’m looking forward to seeing what BSL forum members can do with it. Anyone interested in the Google Doc with the raw data, please comment in the forum.
Shawn began his writing career with Bleacher Report as a Ravens featured columnist and Breaking News Team writer. He moved on to write for Yahoo! and work on the Ravens Central Radio podcast. Most recently, he was an Editor at the Baltimore Wire. Shawn is a 2013 graduate of the Catholic University of America.