Was Keith Smart a Defensive Improvement on Don Nelson?

You ask that question to any member of the Bay Area sports media -- and maybe even any player on the Warriors roster -- and you'd get a resounding "yes." Lead assistant Keith Smart took over for ailing head coach Don Nelson a little more than a week ago. Nelson has apparently recovered from his pneumonia, and Smart confirmed to FanHouse's Chris Tomasson Tuesday that Nelson will be back on the bench Thursday.

Before Nelson called in sick, the Warriors had won four of 12 games. Since, the Smart-led Warriors have gone 2-3. But the one area in which Smart would be perceived to be a massive improvement -- on defense -- the Warriors have not, according to the data, gotten better during Nelson's absence.

This is a tricky matter to unwrap, not the least of which because of rampant sample size issues. Smart coached only five games, and Nelson only 12. This is not a robust data set. But we can take a glance to see if anything sticks out.

To account for the quality of the offenses Golden State faced under each coach, I graded the Warriors' defensive performance relative to the opponent's season-long offensive performance. Doing so gives us a range of results based on the outcome of a game. For example, let's use the case of the season opener. Houston, on the season, averages an offensive rating (points per 100 possessions) of 107.8. In the specific game against Golden State, the Rockets had an offensive rating of 108.2. Houston, then, overperformed on offense just a tiny bit. I did this for each of the 17 games the Warriors have played.

In Nelson's 12 games, the Warrior defense allowed opponents to score an average of 2.8 points per 100 possessions more than would be expected given their season-long performance. In five of those 12 games, the Warriors held the opponent to an offensive rating worse than their current season average. In Smart's five games, Golden State gave up 3.5 points per 100 possessions more than expected. In two games of the five, the Warriors exceeded defensive expectations based on the opponent's offense.

Based on this, Smart's turn at the top seat didn't exactly fix the Warrior defense.

But we can't ignore the one other major defensive component in play: Stephen Jackson. Jackson was traded to Charlotte after the season's first nine games. Nellie, therefore, had Jackson (purported to be an ace defender) available for nine of his 12 games, while Smart didn't have him at all.

But Jackson really doesn't seem like a big defensive aid, according to our numbers. In Jackson's nine games, the Warriors gave up 4.3 more points per 100 possessions than expected. In the eight games since -- five of which were coached by Smart -- the Warriors have given up just 1.6 more points per 100 possessions than expected. Golden State's two best defensive performances, in fact, came after the Jackson trade (a Nellie-coached home win against Portland and a Smart-coached win at Dallas).

Want to make this even more convoluted? Smart has actually run Golden State's defense all season. Nelson has repeatedly stated (almost to a Et tu, Brute? level) that he turned the team defense -- scheme, strategy, practice time -- over to Smart. What Smart has really gained control of during Nelson's absence is the power of player minutes, pep talks and yelling during time-outs. There's murmur that the Warriors play defense with greater ferver under Smart. But the numbers haven't shown it.

In this case, given the limits of the data with the short spans involved, it might be best to just temper enthusiasm for Smart's future, which one would assume involves turning the post-Nellie Warriors into a solid defensive squad as the team's next head coach.

(All raw data, as always, from the invaluable Basketball-Reference.com.)