Scoring defense is a nice stat, but its name is very misleading. It would seem to indicate how many points a defense allows, but thanks to the way the NCAA calculates it, the stat simply tells how many total points the team allows. Things like defensive and special teams scores are left in, making it an inaccurate measure of real defensive performance.
To illustrate the point, I looked at the top six SEC teams in scoring defense from last season. Here's how they ranked with games against I-AA teams excluded:
The presence of three of the conference's four best teams in '08 is not surprising, but the appearance of Tennessee, Auburn, and Vanderbilt would seem to indicate that offensive woes kept them back from winning games they easily could have. You don't know the half of it though.
I ran the same numbers again, only this time I excluded the following:
- Opponent defensive scores
- Opponent special teams scores
- Drives that began in the defense's territory
The first three are obvious, but I wanted to take out that last one too because when an opposing offense starts on the defense's side of the 50, the responsibility for any points scored is just as much or more on the offense or special teams. Here's that chart again without those things listed above:
That says a lot. Special teams breakdowns and the horror of the Clawfense actually added about a touchdown a game to Tennessee's opponents' scoring totals. Ole Miss' indiscretions added nearly the same amount thanks in large part to early season bumbles against Wake Forest (an extra 17 allowed), South Carolina (14), and Florida (14; though the Gators gave 10 right back). On the other end of the spectrum, Florida and Auburn's scoring defense numbers weren't that far off with less than a field goal's worth of difference per game.
Even with these great averages, mostly these teams were not above giving up a decent number of points in a single game. With the adjustments factored in, Auburn still allowed 20+ points four times (maximum in a game of 36). Alabama (max 31), Vanderbilt (max 28), and Tennessee (max 26) all allowed 20+ in three games apiece. Ole Miss (max 27) allowed 20+ in just two games. Florida (max 21) allowed 20+ in just one game: its loss.
Some teams relied on these extra points more than others did. Each one of the six played four of the others, and here's how their offensive scoring averages look with and without the adjustment:
Everyone knew that Florida's opportunistic defense and special teams were good, but against the best defensive teams in the conference, they were worth an entire extra two touchdowns per game. That's an enormous advantage, especially since you can see that Alabama outscored the Gators on a per game basis against these defenses.
We can also see here just how pathetic the offenses for Vanderbilt, Tennessee, and Auburn were against the best defenses. They didn't exactly tear it up before the adjustment, and after the adjustment they were simply terrible. Ole Miss surprisingly wasn't that great either offensively against the best of the conference, though to be fair, three of its four games against these teams came in the first half of the year before they had come into their own.
Now, I know what you're probably thinking. "What about LSU? Didn't Jarrett Lee throw like 12 pick sixes? And how 'bout Georgia? Didn't Stafford throw a bunch of interceptions too?" Well, it was only seven pick sixes by Lee, but yes, I went ahead and did this for LSU and Georgia as well.
Let's start with LSU. The Tigers gave up an average of 30.08 points per game against I-A opponents, an unfathomably high number for an LSU defense. However, that's partially the point of this piece: it wasn't the LSU defense allowing that much, it was the LSU team that did. LSU's adjusted PPG allowed was 17.25, which isn't elite but it is a staggering 12.83 points a game lower than what the team as a whole allowed. LSU's defense did have its struggles though, giving up an honest 38 points to Georgia in one of six occasions of allowing 20+ adjusted points. Against the top six SEC defenses, LSU scored 20.25 a game and an adjusted 16.75 a game for a difference of 3.5 points per.
Now for Georgia. The Bulldogs gave up an average of 24.83 points per game as a team. However, their adjusted points per game allowed was 15.58, a full 9.25 points lower than its team average. That to me suggests that there is a grain of truth in the claim that Georgia might be better off without Matthew Stafford if Joe Cox throws a lot fewer picks. I'm not saying that's 100% true, just that there's at least one piece of evidence to back it up. In total, UGA's defense gave up 20+ adjusted points five times. Against the top six SEC defenses, Georgia scored 21.40 a game and an adjusted 19.60 a game for a difference of 1.8 points per. That shows that for all that the Bulldogs gave up in the way of non-standard points, they were almost completely unable to get some back via big special teams and defensive plays themselves.
Scoring defense is far from the only deceptive statistic that the NCAA keeps. Any and all rushing stats come to mind, since sacks count against rushing yards. However, I hope I've showed one way that you can cut through the mess of scoring defense to get to the truth. After all, Auburn finished a touchdown better than Georgia did against I-A teams in the NCAA's scoring defense category (17.82 versus 24.83), but in my adjusted totals the two programs were dead even (15.36 versus 15.58).
Just one more thing to keep in mind as the season approaches.