Top 10 teams of the last 10 years

Recently ESPN ran a piece listing their top 10 college basketball teams of the last decade.

What troubles me most about this list is not the actual order of the list or the teams on it, but how utterly stupid the reasons used for positioning the teams in such an order are. These are supposedly experts using 'scoring margin' and 'rebounding margin' as useful stats? Scoring margin is obviously influenced by tempo - faster tempo dominant teams tend to win by a larger margin because more possessions means more points. The better stat to use is efficiency margin because it compares a scoring margin that is irrelevant of the tempo the team actually played, fixed at 100 possessions.

Rebounding margin is an even stupider stat to use. Notwithstanding the fact that defensive and offensive rebounding are completely different stats with completely different meanings, rebounding is only correlated with winning insofar as it is affected by or affects field goal percentage. One team often gain a huge defensive rebounding margin with the same defensive rebounding rate if they are shooting a much better percentage than their opponents. Offensive rebounds are important because they often can contribute to an improved field goal percentage by giving teams extra opportunities to score or easy baskets right at the hoop. Of the Final Four teams this year, Georgetown had the highest OR% (a ratio of the number of offensive rebounds gathered to the sum of offensive rebounds and opponents' defensive rebounds) and it, along with a very high 2-point field goal percentage, contributed to their outstandingly efficient offense. Rebounding margin, though, has no meaning. Aren't these guys supposed to be experts?

A couple of notes on the stats I use when comparing teams. Ken Pomeroy is prudent in adjusting raw offensive and defensive efficiencies to strength of schedule in order to come up with an accurate ratings system for a given year. I don't think, though, that adjustments for strength of schedule are necessarily that important when talking about champions or near champions from year to year. Overall NCAA trends do not change much - dispelling, perhaps a little, the myth that some years the game is up and other years the game is down. While I certainly agree that some years the best teams are better than the best teams in other years (otherwise what would be the point of a top 10 best teams list?) but their competition does not vary much from year to year. Some trends have changed over the years, but mainly in the 90s as the 3-point shot became more and more prevalent, but overall things remain rather constant.

Finally, Dean Oliver in his book Basketball on Paper has shown that the 4 most important stats to look at the quality of a team (because of their correlation with winning) are eFG% (3 pointers weighed 50% more since they are worth 1 more point), TO% (ratio of turnovers to total possessions), OR% (described above) and FT Rate (ratio of free throws made to field goal attempts), in that order. Efficiency margin is a consequence of these four stats, and shows a team's ability to bring it all together and perform. I will use primarily these 5 stats in my analysis.

Now on to the list, and my problems with it.

1. 2006-07 Florida (35-5) - A sensible choice, but not inexorable. A lot of this choice is political and heavily influenced by the psychological effect of seeing them play recently. On the whole, though, Florida makes a good case for itself on stats alone. Florida sported an impressive efficiency margin of about 25.5 points and the best eFG% of the teams whose stats I could find (missing a couple until the Mid-Majority comes back online) at 59.6%. However, they were fairly turnover prone (20.9% of possessions ended up as turnovers), were not exceptional on the offensive glass (OR% of 37.66%), and did not convert from the line exceptionally well (FT Rate of 30.2%). The obvious motivating factor making Florida #1 on this list is the fact that it was the 2nd consecutive title. But their resume is lacking compared to some of the other teams for the season: embarrassing losses to LSU, Vanderbilt and Tennessee (blowouts, really) in a 4 game span showed some chinks in the Florida armor. Combine that with less than overwhelming statistical advantages over other teams and it becomes doubly evident how this pick is more or less political.

2. 2004-05 North Carolina (33-4) - This is something I have disagree with, and for much more than the simple fact that it's a Tar Heel team. Actually, as far as UNC teams, I found this team to be rather likable given how much they improved and how much they went through their first two years before becoming a great team in 05. I don't care what team it is, but that sort of thing has to be appealing to any basketball fan. That said, this UNC team is not deserving of 2nd best of the last decade. They had a very good efficiency margin of 23.25 (but it was only 5th best for this period without checking 2000 Cincinnati or 1999 Connecticut yet). They had a 56% eFG%, a very good result but again, 5th best of these teams. They had the second worst TO% of champions, not an exceptional offensive rebounding team (OR% of 39.5%) and a very good FT Rate of 32%. UNC had balance, though, and it was a tough team to beat. I don't really understand, though, how they consider a team who had such obvious blemishes on their record (losses to Santa Clara and failure to win conference tournament), combined with no reason to believe that this team was as statistically dominant as other teams of the last decade to be as good vis-a-vis other champions.

3. 1998-99 Connecticut (34-2) - Since UConn's official website keeps paltry stats for its own teams and since I have been unable to get my hands on a full record of team statistics for the year, I'm really unable to comment on this team too deeply. I think this is a very merited position for this UConn team if only for the fact that they beat the, statistically, most dominant team of the decade in the championship game. They also had very few blemishes on their record, and excellent players. When I know more, I can comment more, though.

4. 2003-04 Connecticut (33-6) - Quite an unusual pick for a team that wasn't even a #1 seed in the tournament. They were certainly one of the better defensive teams of the last decade (90.9 defensive efficiency), but (and this is something to be saved for another entry) defense alone does not win championships. Their efficiency margin was not spectacular (21.2) and they were awful at converting from the line (21.5% FT Rate). They had just a 53.2% eFG% (for reference, Duke had 52.5% this year and we were 22-11), but were good with the ball (19.4% TO%) and were a pretty good offensive rebounding team (OR% of 41.3). This Connecticut team had great front-court depth, but were not even close to being one of the most dominant teams of the decade. It took a late rally against a very good, but not great, Duke team in the Final Four to even make the finals. The best teams don't leave so little room for themselves.

5. 2000-01 Duke (35-4) - I will not deny my allegiances here, but this is a patently incorrect pick. Duke 2001 can reasonably be placed as low as 3 or as high as even 1, but #5 is a rather insane position. Of all teams I looked at in this period, this Duke team had the 2nd highest efficiency margin (26.77) and was one of just two teams to score over 1.2 points per possession. The article concedes that this is a Duke team who lost to only very good teams and by very small margins (1 point loss to Stanford, 2 point losses to Virginia and UNC, and an 11 point loss to Maryland at home during the game when Boozer broke his foot), and who dominated their NCAA tournament opposition (winning every game by double digits). Duke had a higher eFG% than any team except Florida 06 and 07, and Duke 99 at 56.2%, and the lowest TO% of the last decade of champions at 18%. Furthermore, the article criticizes Duke for not having a huge rebounding edge - but who cares? The most important rebounding stat correlated with winning (OR%) is not low by any means (37.8%; higher than both Florida teams, and 02 Maryland). The article also criticizes Duke for having free throw troubles, which is probably the only criticism you can make of this team. Their FT Rate was only 27.6% and there were stretches where they struggled from the line (costing us at least 1 game). However, as far as criticisms go, this one is a little odd to drop a team who deserves a really high ranking so low. A high turnover rate is a much worse liability on a team, as it is more closely correlated with losing than a lower FT Rate.

6. 1999-00 Michigan State (32-7) - Sensible for its position, but a deeply flawed team. They were the best defensive team, statistically, of all champions (just an 89.0 defensive efficiency), and the best offensive rebounding team of all champions (42.1% OR%). This is a team that made a living grinding it out and crashing the boards hard, but they had difficulties against teams with more efficient offenses. The 1999 version of this MSU team had nearly identical stats, and was an even better offensive rebounding team, but went 0-3 against UConn and Duke that year (both very efficient offensive teams). Simply put, their offense was not strong enough to compete with some of the better offensive teams. They had just a 53.3% eFG%, and the highest TO% of champions at 22.0%. They were also rather poor at converting from the FT line as they had a 27.9% FT Rate. My suspicion is head-to-head with some of the better offensive teams on this list, MSU would struggle to keep up.

7. 1998-99 Duke (37-2) - I understand perfectly why this team is at this position. Winning a championship means a lot, but I would be remiss not to glow a little over the accomplishments of this Duke team. As Joe Lunardi points out, Duke from 99 is #1 on paper - and the margin is not that close. Most statisticians have concluded that there is no ascertainable factor that determines close games. Most chalk it up to some intangible "luck" notion now. After all, one or two possessions throughout a game can mean the difference between a win or a loss in a very tight game. This team lost two games by a total of 5 points all year and sported (by a large margin) the highest efficiency margin of any team in question (33.3). They had the highest offensive efficiency (124.0), and only 04 UConn and 00 Michigan State sported a more efficient defense (Duke's DE was 90.65). They had the 3rd highest eFG% at 57.4, a fairly low TO% at 19.6%, a better OR% than any champion (44.4%), and the best FT Rate of any team I looked at in the period (33.0%). So, again, in terms of performance, no team in the last decade was as good as Duke in 99. They had a rare combination of size, athleticism, depth, defensive intensity, desire and experience - but had a bad game at the wrong time and came up short. History forgets the losers of title games, unfortunately. 99 Duke will go down with 91 UNLV as probably the two best teams to fail to win the championship. I suppose the ranking at this position is just, but if this were really a determination of which team was actually the best in the last 10 years regardless of championship or not, Duke 99 is it without question.

8. 2001-02 Maryland (32-4) - A just position for a very good, but not great, team. Maryland was hard to beat, because of experience, balance, good coaching and an excellent leader in Juan Dixon, but they lacked the overwhelming talent that some of the greater teams (Duke 01, Florida 07, UNC 05) had. They had a small efficiency margin (18.7), with neither an overwhelming offense nor defense. They did not shoot particularly well from the field (53% eFG), nor were they very good on the offensive glass (just 36.0% OR%), but they converted from the line well (30.1% FT Rate), and, most important to their consistency and success, took care of the ball (just 18.2% TO rate). I think I would have placed them lower, perhaps 9 or 10, and below both non-champions who deserve a position on the list (05 Illinois to join 99 Duke).

9. 2004-05 Illinois (37-2) - A real interesting team. Illinois was severely lacking on the interior, which explains how Sean May dismantled them in the 2005 championship game, but they were incredibly careful with the ball, turning it over on just 16.7% of possessions. Combined with a very good eFG% at 56%, Illinois managed a very good offensive efficiency (116.8) despite having a tremendously low OR% (34.3) and an abysmal FT Rate (21.0%). Their efficiency margin was a very good 24.12 (actually better than UNC 05's). This team justifiably belongs above 2 and possibly 3 champions of the last decade, but the other 7 champions and Duke 99 would manhandle this Illinois team on the inside. I'd put them at #8 on my list above 02 Maryland and 98 Kentucky.

10. 2002-03 Syracuse (30-5) - How can this team be above 98 Kentucky, or 00 Cincinnati? 'Cuse had the lowest efficiency margin (13.8), and the lowest eFG (51.9%). The 2003 season really lacked a dominant force of a team, so it's not surprisingly that a 3 seed would come out of no where and win it all, but they did not do so in a dominating fashion, nor did they exhibit any level of dominance during the regular season. I would leave them off the list myself.

My list:
1. Florida 2007 - Like I said, the political pick. Going back-to-back is really hard in itself, and they should get credit for it even if the other stats don't sway their way.
2. Duke 2001 - I made the case earlier and I stand by it.
3. UConn 1999 - I can stand by this argument later more once I have the stats, but beating 99 Duke counts for a lot in my book.
4. UNC 2005 - There's lots to like about this team, but we can't let the recency factor overwhelm our rational faculties.
5. UConn 2004 - Excellent defensively, numbers for the season are probably artificially deflated due to Okafor's injury.
6. Duke 1999 - Best on paper, probably the best team but doesn't deserve a higher position because they choked the last game.
7. Michigan State 2000 - Excellent on the glass and defensively, but not spectacular offensively.
8. Illinois 2005 - Probably best guard-play of all teams here, but not balanced. Very good offensively, though, despite interior absences.
9. Maryland 2002 - Very good, but not great team. Poised, experienced, balanced - took very good care of the ball, but not overwhelmingly talented. Didn't blow teams out or have spectacular numbers.
10. Kentucky 1998 - Not even in Honorable mention for ESPN's list, this Kentucky team had very similar stats to Maryland, but was better on the offensive glass. The style they played was risky, but allowed them to come back from big deficits better than any team on the list except Duke 01. Cincinnati 00 might deserve this position more but fate excludes them by shattering K-Mart's leg...