I was responding to this post of yours, which was a response to Rob.
"It also shows the level of play that this team is capable of, and the level that they should be playing at. For the most part, two-thirds of the way through the season, and this team has greatly underperformed their actual level."
"It" seems to refer to the second half of the Virginia game; it was the "it" in the post with which you were agreeing.
I say that the second half of the Virginia game tells us very little on its own about the team's defensive ability. Let's go back to the exam analogy. It is usually the case that students understand some topics better than others; for instance, I would have students who understood ordinary fiduciary duties very well but compensation-related issues somewhat less well. All my exams would feature ordinary fiduciary duties (that's the core of the course); sometimes I would include compensation but not always. The student who doesn't really get the compensation issues is not going to be the best corporate law student in the class; but if compensation isn't on the exam, you wouldn't tell from the exam score. If UVa was not running the actions that give us trouble, or not running them well, you'd see a good performance against them and it would tell you little about the overall ability.
To complete the analogy: If my exam was a complete sampling of course topics, then it would be more of an exhaustive measure. But it wasn't, because it couldn't be. Same is true with a single half of basketball. It was one half, of one game, against one team, presumably playing at a certain quality level. That doesn't tell you much about overall performance. It *might* tell you that the first half defense was below ability, but even that is iffy because there's another variable in the equation: the quality of the offensive execution.
In general, I think it is very difficult to distinguish offensive execution from defensive execution. The best way to do so is look at a large sample of games, not just minutes played. Even that isn't necessarily going to tell you how a team performed on a given day, but whatever. Which is why I think the reliance on a much larger sample size to consider the team's ability is correct, and looking at a single game as evidence of the "ability" is unfair.