People criticize the 2005 SC defense, and I get that to some degree. But I think they were quite a bit better than their perception. Mainly, I think it's that they weren't as good as 2004, or what the next group rounded into by the end of 2006 (and that same group's 2007 and 2008 seasons, especially). But those other defenses were stellar, and 2005 shouldn't be thought of as poor simply because they weren't as good as the units that preceded and succeeded them.
I'll say this, if Jayden Daniels had 2005 USC's defense, then 2023 LSU would have been runaway national champions and nobody would've even gotten close.
They probably would've been top 15 if we'd had advanced stats like FEI, SP+, F+ back then. Texas scored 38 on that unit.
I guess this technically goes on the Big Ten board ('cuz USC is now a B10 team but I will never accept that in my mind). But Texas is now an SEC team, so this sort of goes here (but I also don't accept Texas as an SEC team. Or A&M. Or Arkansas. Missouri, I'm pretty sure is an ongoing hallucination of the conference).
I checked the bottom line numbers just to see what scores USC allowed that year. Not counting Texas, they allowed 21.3 ppg. Not bad, for an offensive league. Of those 11 games that weren't Texas or Fresno (more on them in a second), Notre Dame (31) and ASU (28) were the teams that scored the most. I don't remember Notre Dame as well that season, other than the infamous Bush-Push ending, but since ASU played LSU that season I remember them well, following them the rest of the way. ASU was a high-powered, dangerous offense, and just ask LSU's slew of future NFL defensive stars about them. Fresno was the lone team to actually score more than Texas (42), but that score is a clear outlier in what USC's '05 defense allowed, and frankly.....it was Fresno State, and those games happen. Like many good statistical models, I'm more than willing to remove outliers for analysis. Fresno, Notre Dame, and Arizona State inflate that average, and outside of that SC was giving up just 17.2 ppg. I understand any average improves when you remove the worst cases, but understanding the skew and kurtosis of a distribution is important analytical context. In fact, for another data point SC shared with LSU, Arkansas also played both teams and scored exactly 17 on both. While SC is popularly thought to have been "down" on defense that year, LSU was perceived--correctly, imo--to be quite good.
My point is SC's defense was nowhere near as meh as the narrative always seemed to claim. Texas scored
41 (ignore my mistake in the quote) when that defense was trying its hardest, had a month to prepare, and was, frankly, much better than its perception. Whether that was because they were overshadowed by their offense, not quite as good as the impression that the dominant display the OU NC game left everyone with, a combination, or something else entirely, I'm not sure.
Likewise, the Texas defense didn't allow anything close to 38 during the year. They allowed just 14.6 ppg. They also have "outliers" with Okie. St. (28) and A&M (29).
Texas clearly had the better defense, but SC's was quite good. The 41-38 score was quite a testament to how good those offenses were.
That's a long-winded explanation of why that game was the best non-LSU NC I've ever seen, and how, even despite the higher score, a smothering defense-lover like me could not reasonably nitpick it.
