Monday, March 19, 2007

Year of Balance

Last year at this time, I looked at how the offensive and defensive efficiencies of the remaining teams stacked up, and concluded it was offense-heavy. Then, three defense-heavy teams (and Florida, third best in terms of balance, behind Texas and UConn) stormed into the final four. This year, the meme seems to be balance. Though the offensive side is a little more top-heavy, a tidily symmetrical 9 of the top 16 offenses and 9 of the top 16 defenses remain in the field, with only one severe outlier (SIU's offense). "Specialist" teams such as Illinois (#2 defense, #111 offense), Texas (#5 offense, #62 defense), and Boston College (#10 offense, #92 defense) have already been shown the door. In this year of balance, only two teams rank in the top 10 in both offense and defense - UNC and Ohio State. Here's how the remaining squads stack up:

1) UNC - 124.2
2) Florida - 123.7
3) Georgetown - 123.6
4) Ohio State - 120.9
6) Texas A&M - 120.4
8) Oregon - 119.7
12) Kansas - 119.2
13) Pittsburgh - 118.9
16) Butler - 117.7
21) Tennessee - 117.0
22) Vanderbilt - 116.7
24) UCLA - 116.7
28) Memphis - 115.7
37) USC - 114.2
38) UNLV - 114.1
81) Southern Illinois - 109.2

1) Kansas - 81.8
3) UCLA - 84.8
4) UNC - 85.0
5) Memphis - 85.6
10) Ohio State - 86.3
11) Texas A&M - 86.3
12) Southern Illinois - 86.7
13) Florida - 86.9
16) Georgetown - 88.3
25) USC - 90.4
26) Pittsburgh - 90.5
45) UNLV - 92.5
47) Butler - 92.7
50) Tennessee - 93.1
51) Oregon - 93.2
56) Vanderbilt - 93.9


Anonymous said...

You must be mistaken, Paul. I have heard analyst after analyst after talking head tell listeners that UNC does not play any defense to speak of--when Carolina wins it is due to its offensive firepower. But what you have written here makes it sound as though Carolina is actually competent at defense.

You might want to re-check your numbers. I doubt that all of those paid commentators could be wrong.

Unknown said...

UNC has one of the best defenses in the country. I suspect that the paid commentators to whom you are referring fall victim to the trap of the unsophisticated, namely conflating the quality of a team's defense with the average number of points that team allows per game.

This is particularly problematic with a team such as UNC that tends to play high-tempo games. The more possessions a team has, the more opportunities to score, and inevitably the more it will score. This tempo is not an indictment of UNC's defense, however.

The same thing occurred when all of those analysts harped on how apparently awful Duke's offense was because it ranked last in the ACC in scoring average. Duke also played the slowest tempo in the league, making its offense look worse than it was (and its defense look better than it was) to the unsophisticated analyst.

Michael said...

UNC can have one of the best defenses in the country... but their problem is the occasional lapse in defensive intensity. The reason I am skeptical of their ability to win the title is not in their talent, or the aggregate defensive or offensive performance of their team over the course of the season, but rather, because of their tendency to lapse on defense and make mental errors attributable to youth.

Paul Rugani said...

I salute all three commentators, all of whom are correct (and I enjoy the snarkiness of the first - sarcasm will always be welcome here).

UNC is more vulnerable to defensive lapses than offensive - only 2 sub-average offensive games, but 13 sub-average defensive ones. Moreover, all 6 of the Heels losses have come with the opponent scoring a point-per-possession. This is a good count along game at home - if USC is scoring more than one point per possession, they have a 50% shot of winning.

I'll have a little more on this in the next major post on the blog, which will preview UNC and USC (both of whom I've seen a lot this year).

Anonymous said...

Interesting point about the variablity of offensive and defensive results as opposed to the mean results. Some teams may be excellent but "bi-polar" if you will, a trait that is less likely to be costly in a seven game series than in a single-elimination tournament.

If it were possible to bet on a given team not to win the tournament, this would be similar to an option to sell short in financial theory and the value of the option would increase with respect to the variability of the expected outcome.

Essentially what that means as a practical matter is that it is necessary, but not sufficient to be an excellent team to win six games straight.

The best team may not even usually win, but it is rare to have a low seed win it all, because for a top team, just one performance below the norm can boot them out of the tournament, but for a weaker team, it takes several outlier outcomes to win it all. George Mason might have gotten two or three last year, but they were still a long way away.

The 1982 Carolina team which was named the greatest of all time by ESPN (the greatest team actually is probably the 1999 Blue Devils, but you had to win it all to be considered), won its NCAA tourney games by the following scores: 52-50, 74-69, 70-60, 68-63 and 63-62.

They were far from spectacular in most of that tournament, but they were steady and not prone to lapses and the team that made the last mistake that year lost.

But when you compare them to Houston in 1983, well, Houston was spectacular all through the tournament until the final game when they weren't and the headier team, NCSU won.

Houston's mean performance was almost certainly better than UNC's the year before but the NCAA tourney is not the NBA, where one game means relatively little.

I wonder if you can break you information out to home, away and neutral courts, as young teams like UNC might be underperformers more often in hostile environments.

East Rutherford should be fairly neutral since it is featuring Georgetown, which is closer than UNC, although the New York area has often rooted for Duke or UNC.