When comparing offenses and defenses, most yardage stats for a team are given in yards per game. It's assumed that good teams gain a lot of yards per game and yield few. But in some cases, looking at yards per game is misleading, and looking instead at yards per attempt is more instructive.
Take Wake Forest, for example. In a 13-10 loss last Saturday they held Navy to zero passing yards, and accordingly their average passing yards yielded per game went down. However, Navy didn't even attempt a single pass the entire game! Everyone knows Navy doesn't pass much, and they only average 77.25 yards per game. But even when adjusted for opposition, Wake Forest's passing defense stats improve: last week they were #101 in per-game pass defense, and this week they're #68, for "holding" Navy to 0 yards.
That's why yards per attempt statistics are important. By that measure, neither Navy's passing offense nor Wake Forest's passing defense are drastically affected when Navy doesn't attempt a pass.
Lately I've created four new stats pages: for Rushing Offense Per Carry, Rushing Defense Per Carry, Passing Offense Per Attempt, and Pass Defense Per Attempt. All are adjusted, of course, for opposition, and give a more complete view of a team's offense and defense. But they aren't without their own misleading aspects as well. Here's a rundown of each, compared to its per-game equivalent.
Here are the top five teams in Per-Game Rushing Defense & their rank in Per-Carry Rushing Defense
Rank Team Rec YPG [rank] YPC[Rank]
1. Alabama 8-0 48.2[ 1] 2.1[ 1]
2. Ohio State 6-2 65.9[ 2] 2.4[ 3]
3. Texas 7-0 69.1[ 3] 2.4[ 3]
4. West Virginia 6-1 69.9[ 4] 2.5[ 5]
5. Wisconsin 5-2 78.8[ 5] 2.8[ 15]
For rushing defense, there isn't a lot of difference between per-game and per-carry average. Teams with good defenses will do well on both measures, as there are no odd circumstances that cause a schizm. Since both are corrected for their opponents' rushing ability, the lists are fairly similar.
But when we look from the offensive perspective, things are a bit different. Teams like Georgia Tech, Navy, and Air Force—all of whom run the triple option and run the ball 90% of the time—tend to dominate the per-game rushing stats for logical reasons. But are they really the best at running the ball, just because they carry the ball 80 times?
Rank Team Rec YPG[rank] YPC[rank]
1. Nevada 4-3 329.5[ 1] 7.1[ 1]
2. Georgia Tech 7-1 325.6[ 2] 5.6[ 7]
3. Navy 6-2 286.5[ 3] 4.5[ 29]
4. Florida 7-0 285.4[ 4] 6.1[ 3]
5. Air Force 4-4 272.0[ 5] 4.4[ 37]
While it's no surprise that Florida registers well both in game and per-carry average, Nevada coming out #1 in both measures shows how strong they are on the ground. And Georgia Tech deserves its rep as a great running team. But the service academies fall short; while they run so often their per-game averages are great, per-carry they are only above-average to good. Without seeing the per-game averages you don't get a full picture of the Navy and Air Force offense.
As in rushing, the teams that pass the most don't necessarily have the best averages:
Rank Team Rec YPG[rank] YPA[rank]
1. Texas Tech 5-3 426.8[ 1] 8.0[ 27]
2. Houston 6-1 394.5[ 2] 7.3[ 48]
3. Hawaii 2-5 354.5[ 3] 8.1[ 26]
4. Bowling Green 3-5 347.5[ 4] 7.1[ 59]
5. Kansas 5-2 346.2[ 5] 7.6[ 38]
A common thread among these teams is some type of spread offense, and an emphasis on quick, short passes. This naturally keeps the average down, and makes comparing teams by average yardage less informative about the team's passing game.
Not only that, but the same teams that cause the rushing figures to be disparate also cause some confusion with the per-attempt passing averages: these teams pass so few times per game, they actually end up with a very high per-attempt score. Here are the top five in Yards Per Attempt:
Rank Team Rec YPA[rank] YPG[rank]
1. Georgia Tech 7-1 12.1[ 1] 153.8
2. Virginia Tech 5-2 10.7[ 2] 198.8[ 77]
3. TCU 7-0 9.9[ 3] 246.2[ 34]
4. Miami FL 5-2 9.7[ 4] 277.9[ 19]
5. Navy 6-2 9.4[ 5] 69.1
These results are even more radical! Georgia Tech and Navy, two of the "worst" passing teams in the country, rank as two of the best per-attempt. Maybe if they passed more they could dominate? Clearly, the surprise factor is at work here. Teams key on the run against Navy, Air Force, Georgia Tech, et al, so it's not surprising that the few times they throw, they tend to go for it all. And sometimes succeed. There's little correlation between a team's per-attempt average and the quality of its passing game, so due to this and the other issue above, it's best to reference per-game statistics if you want to see which teams have good passing games.
But on defense, the per-attempt stats are necessary. This is due to of one of the per-game stats' biggest flaws: teams with poor rushing defense often grade as teams with great passing defense—simply because no team has to bother with putting the ball in the air if they can get yards on the ground.
Rank Team Rec YPG[rank] YPA[rank]
1. Florida 7-0 109.7[ 1] 5.0[ 5]
2. Nebraska 4-3 128.3[ 2] 5.4[ 13]
3. South Carolina 6-2 128.3[ 3] 5.6[ 18]
4. Boise St 7-0 138.8[ 4] 4.9[ 4]
5. Texas 7-0 149.9[ 5] 5.6[ 16]
So far, no problem: the teams that look like the best per-game are solid, and rate well per-attempt, too. But look a little further down the list and you see teams like this:
13. Bowling Green 3-5 163.6[ 13] 6.8[ 49]
14. Eastern Michigan 0-7 163.7[ 14] 7.5[ 84]
Unlike the teams above, Bowling Green and Eastern Michigan have terrible rushing defenses, ranking 116th and 120th. Do they really have top 25 passing defenses to go with that? It's possible, but it's also likely that opponents are playing the "why pass when we can run so easily?" game, passing for a small amount during the game, and making the pass defense look good on a per-game basis. And looking at the per-attempt defensive stats, this seems to be true. Bowling Green isn't terrible at #49, but it's a lot less appealing than #13. And Eastern Michigan is pretty bad in reality, ranking #84 instead of #14. Passing defense is one area where per-attempt stats are necessary for confirmation that a team is actually good against the pass.
The most informative per-attempt stats are Pass Defense - Yards Per Attempt, and Rushing Offense - Yards Per Carry. Both give important additional information that the per-game figures don't. The former works to expose teams that have great per-game averages but just aren't being tested. The latter shows which teams really have the best rushing offenses, independent of the fact that some teams run almost exclusively.
Rushing Defense - Yards Per Carry mostly confirms what we know from the per-game stats, but it can be an important second reference.
Finally, Passing Offense - Yards Per Attempt can be a misleading statistic for a couple of reasons. One, teams with spread attacks that pass often and for short yardage end up looking worse, while teams that rarely attempt a pass can have outsized averages simply based on small sample, and on the "surprise factor" that makes their passing game look powerful when in reality they just catch teams off guard.