Ziller compares each opponent's overall offensive efficiency to their offensive efficiency in their game(s) against us. This is a fine shorthand approach, but it ignores the type of offense that each team brings to the table. We are more defenseless against certain types of attacks than against others; the thing that has most killed us so far, of course, is size. Between the injuries to Biedrins and Turiaf, the limited minutes of Randolph, the heavy use of The Worst Starter In Basketball and the frequent presence of Maggette and Radmanovic at the four, we are aberrantly easy pickings for any team with decent bigs. We'll do worse against a big poor-scoring team than against a small good-scoring team.
Therefore, you can't simply ask, "How good were our opponents on offense?" You also have to ask, "How big were our opponents?" And when you ask that question, you see that our opponents in Nellie's twelve game were, on average, much smaller than our opponents in Smart's five games. Nellie faced the undersized Rockets, the undersized Suns, the Love-less Wolves, the Knicks, the Cavs without Shaq and Varejao... he faced a bunch of small teams. Smart's opponents were the Mavs, the Spurs, the Lakers, the Pacers and the Nuggets, all teams with strong frontcourts. (Nellie faced the Pacers too, but they didn't have Troy Murphy then. I know, I know.) In context, Smart had a far more difficult defensive task than Nellie, and a far more difficult defensive task than Ziller estimates. We are unforgivably small and weak up and front, and Smart wasn't lucky enough to face teams that couldn't exploit that.
Rest is also a huge issue here. Of Nellie's twelve games, only three were the second halves of back-to-backs, and one was at home against the Wolves, my pick for the actual worst team in the NBA. Two of Smart's five games were the second halves of back-to-backs, and both on the road: in San Antonio the day after playing in Dallas, and the laughably impossible Denver visit the day after playing in Oracle. You're not likely to get energetic defensive efforts out of a running, short-handed roster in situations like that. The Warriors were much more tired on average during Smart's tenure; that has to be accounted for.
Finally, there's player availability. Ziller correctly points out that Stephen Jackson's departure probably doesn't skew anything in Nellie's favor much, as there's just no evidence that he's been a defensive asset lately. However, Ziller ignores the departure of another player who was a real defensive asset: Kelenna Azubuike, by far our best defender early on. Nellie got eight games of 'Buike, Smart got zero. Big, big difference. Nellie also had Biedrins for a couple games... Biedrins wasn't any sort of world-beater in those games, but we're a much more reasonable approximation of an NBA team when he's out there.
I don't think Keith Smart becoming our head coach would be any sort of panacea, and our improved recent play may have more to do with Jack's departure than anything on the coaching side. But considering the context, we did play better, on both sides of the ball, under Smart. And while I have no idea if he'd be a good head coach, I'm damn sure he'd be better than the eyes-glazed-over Nellie we've been subjected to for the last year-plus.