Last offseason, I argued that the Cowboys should use pre-snap motion less often. In my 2009 study on Cowboys motions, I found the offense was generally less successful on motion plays, averaging 0.7 less yards-per-pass and a full yard less per run. As a result of those findings, I wrote:
Garrett should steadily decrease the motion rate until the defense compensates enough that the Cowboys’ yards-per-play reaches its peak. My guess is that this is around 25.0 percent. At this point, it is likely the rate of big plays and negative plays will also be maximized and minimized, respectively, creating situations of generally optimal efficiency for the Cowboys’ offense.
Garrett used motion on 42.5 percent of snaps in 2009. As I suggested, that rate dipped this past season. Here are the Cowboys’ 2010 motion numbers. . .
Right off the bat, you can see the Cowboys’ motion rate dropped to 34.4 percent–not quite the 25.0 percent I suggested, but still an improvement. With that fall came an increase in efficiency, at least in the passing game. The Cowboys averaged over 0.8 “extra” yards-per-pass on motion plays and garnered a higher rate of big plays (10+ yards)–31.3 percent versus 22.3 percent on non-motion plays.
Once again, the Cowboys motioned on a higher rate of run plays than pass plays. The 175:176 ratio is incredible, and those 175 runs represent 45.8 percent of all Cowboys’ runs–very similar to the team’s 48.9 percent rate from 2009. Note that the “actual” run totals are slightly skewed because I count only “called” runs, not quarterback scrambles, kneel downs, etc.
The reason for the increased motion rate on run plays seems simple enough; the Cowboys frequently remain static pre-snap in situations where the defense knows they are going to pass. For example, Dallas motioned on only 12 of the 197 plays they lined up in “Gun TE Spread” (left)–that’s a rate of just 6.1 percent, down from 12.5 percent in 2009.
Thus, while the run/pass ratio after motions is a bit “off,” it creates no real competitive advantage for the defense.
So we know the Cowboys motion more on run plays, but is there any causation behind this correlation? My initial thought was that the the drop in yards-per-rush was caused by a possible tendency to motion on short-yardage plays. Thus, the upside would be limited and the averages would suffer.
However, on short-yardage plays (which I defined as three yards-to-go or less), the Cowboys motioned just 12 times out of a possible 97–only 12.4 percent of the time. That’s far below the overall rate of 34.4 percent. Thus, while it is good that Garrett effectively spreads out motions among various downs and distances, the low yards-per-rush on motions cannot be attributed to an abundance of short-yardage plays.
Another possibility for the lack of success on motion runs is predictability. After watching as much film as I do, there are times when I can predict with great precision what play the Cowboys are going to run. How and where they motion is a big factor in my ability to do this. Dallas will frequently motion the fullback to the play-side just before the snap, for example. Only rarely does the fullback motion to the side of the formation opposite the play-call.
If I can read these tendencies, NFL defenses can do it.
Of course, the Cowboys simply need to run the ball more effectively on all types of plays. They weren’t particularly dominant on non-motion runs either.
I love that I saw a decrease in motions from Garrett and the ‘Boys in 2010. The Cowboys found much more success via the pass on plays which invoked a pre-snap motion, and, although the team’s run efficiency plummeted, the relationship between motion and non-motion runs actually converged.
We also saw a greater rate of big pass plays and a lower sack rate (by far) on motion plays as compared to 2009. The rate of total negative plays remained steady at 11.1 percent.
Ultimately, I still think 25 percent is the magic number. If the Cowboys can motion around one-in-four plays in 2011, I think they’ll see an increase in overall offensive efficiency.
Of course, regression to the mean tells us we’ll probably see that anyway.