Thursday, January 14, 2010

Picking Winners & Losers


I Really Do Like Luke Winn's Analysis, But...
Several years ago Sports Illustrated's Luke Winn observed that virtually (but not actually) all teams that made the Elite Eight had (adjusted) defensive efficiencies (per Ken Pomeroy's Stats Page) -- defined as points allowed per 100 possessions -- that ranked them among Pomeroy's top 25/50 in Division 1. And sure enough, according to the tables he produced (which are 100% accurate by the way), just as he claimed, "...In the six years kenpom.com has tracked tempo-free statistics, no Final Four team has been ranked outside the top 25 in adjusted defensive efficiency (a figure that factors in level of competition). And only two Elite Eight teams in the past six years have ranked outside the top 50 in adjusted defensive efficiency. Having a decent offense matters, too..." (as he phrased it for the annual circa New Year's column). All of this is interesting, in a self-fulfilling prophecy sort of way, but can "knowing" that the Final Four teams will have adjusted defensive efficiencies that rank them in the top 25 be of use now?

Critics of the traditional (AP & UPI/ESPN/USA Today/Coaches, etc) weekly polls have cited the pollster's inclination to give disproportionate weight/value to a team's name/reputation, conference affiliation and sometimes overvalue (or undervalue) the team's win-loss record when developing their ballots. Pomeroy's own Ranking System, based entirely on performance-based statistics, shows an ordered list that differs with the weekly polls over the ranking of a number of well-regarded D1 teams. Winn's intent, well stated (but unfortunately easily conflated by fans), is to identify those well-regarded (ie highly-ranked) teams whose performance-based statistics may undermine their (sometimes very gaudy) win-loss record. Winn "does it right". He draws up a list of highly ranked (in the late OOC polls) teams whose defensive efficiency, at that point in the season, are not consistent with efficiency profiles of the teams that have historically made the Final Four/Elite Eight. It is, unfortunately a too short leap from what Winn proposes, to what the fans conclude -- that a team which does not meet the historic criteria in early January will not meet the criteria by season's end (or Selection Sunday). I have reproduced Winn's table, for seasons 2006, 2007, 2008 & 2009, but I have added the (adjusted) efficiency rankings for those Elite Eight/Final Four teams in early January (about when Winn typically penned his analysis), and again in early February, about the mid point for most conference regular seasons (and about 75% through the season).


Season-endEarly FebEarly Jan
FinOff AdjDef AdjOff AdjDef AdjOff AdjDef Adj
2008-09
North Carolina NC 124.2 (1)89.6 (16)121.9 (1)88.1 (17)123.5 (2)85.8 (18)
Michigan State RU 115.0 (20) 88.4 (10)115.9 (12)92.7 (45)113.8 (30)92.1 (59)
UConn FF 116.6 (15)84.8 (3)118.8 (5)85.7 (7)120.2 (5)87.4 (26)
Villanova FF 115.0 (22) 89.6 (15)111.5 (36)87.3 (12)108.8 (59)86.8 (25)
Pitt EE 122.2 (2)92.0 (35)121.7 (2)90.2 (25)117.9 (12)84.8 (12)
Missouri EE 117.8 (8) 89.1 (13)116.7 (8)89.9 (23)118.8 (8)87.6 (27)
Louisville EE 113.4 (31) 84.2 (2)106.5 (94)81.1 (2)105.1 (99)78.4 (2)
OklahomaEE 118.3 (6)92.1 (36)119.7 (4)92.3 (44)117.9 (11)90.6 (47)
2007-08
Kansas NC 125.3 (2)82.8 (1)123.1 (1)81.4 (3)118.2 (10)80.1 (2)
Memphis RU 121.3 (4) 83.9 (4)114.2 (32)78.9 (1)117.3 (15)80.6 (3)
UCLA FF 119.7 (7)83.9 (3)122.4 (3)85.5 (8)116.8 (18)83.9 (7)
North Carolina FF 126.0 (1)89.4 (19)122.3 (4)89.9 (33)122.4 (3)89.6 (38)
Louisville EE 115.2 (32)84.1 (5)113.5 (35)85.2 (7)109.3 (65)86.8 (17)
Texas EE 123.8 (3)91.8 (36)122.9 (2)97.0 (89)123.8 (2)93.4 (66)
Davidson EE 117.7 (14)91.3 (31)109.9 (60)94.1 (65)109.4 (63)97.4 (116)
Xavier EE 118.4 (9) 91.7 (35)119.1 (8)89.4 (25)124.1 (1)90.4 (40)
2006-07
Florida NC 125.4 (1)87.4 (12)124.6 (1)85.7 (11)121.6 (5)82.7 (9)
Ohio State RU 123.7 (4) 87.9 (15)122.8 (4)86.3 (13)121.5 (6)87.1 (23)
Georgetown FF 124.8 (2)89.3 (20)123.4 (3)90.7 (36)118.2 (13)87.0 (22)
UCLA FF 116.8 (23)84.0 (2)118.1 (14)82.4 (4)120.2 (9)83.7 (11)
Oregon EE120.3 (8) 93.4 (51)117.5 (18)93.9 (63)114.5 (27)88.8 (31)
North Carolina EE 123.9 (3) 85.6 (4)121.1 (7)80.6 (2)123.7 (3)81.8 (7)
Memphis EE 116.6 (25)86.9 (11)116.7 (22)85.6 (10)113.2 (40)82.6 (8)
Kansas EE 117.8 (17) 82.2 (1)113.1 (41)81.9 (3)113.0 (41)81.1 (4)
2005-06
Florida NC 119.4 (2)87.2 (5)114.7 (13)88.9 (22)114.9 (14)87.3 (28)
UCLA RU 113.0 (28)85.1 (3)109.9 (43)88.9 (23)113.8 (19)93.8 (88)
LSU FF 109.9 (50) 85.7 (4)112.2 (23)87.4 (14)108.9 (49)86.0 (17)
George Mason FF 110.0 (49) 90.0 (18)107.5 (66)87.3 (15)105.8 (75)88.8 (34)
Memphis EE 112.9 (29) 87.4 (6)112.1 (24)86.3 (8)109.5 (44)80.4 (8)
Texas EE118.8 (4)88.7 (10)115.5 (8)80.2 (1)115.1 (13)82.8 (5)
UConn EE 119.2 (3)89.9 (16)114.6 (14)84.5 (5)116.0 (10)87.5 (29)
Villanova EE117.0 (11)89.9 (17)118.4 (2)88.5 (19)121.2 (2)83.3 (6)

I have marked in orange where the EE/FF team did not meet the criteria, and in yellow where the team is within 1-2 spots away from the limits observed by Winn (Top 50 for EE teams, Top 25 for FF teams). While Winn's original article (written in December of 2005 I believe, but I have been unable to locate it) suggested a potent offense was the key. A look at the 2006 Final Four teams blows that theory up, but note how consistently in 2007 and 2008, that top 5 offensive teams did indeed make the Final Four. I have also highlighted using green, those offensive rankings that fall outside of the top 30. The criteria is clearly not iron-clad; there are a lot of exceptions. Teams can progress (and regress) as the season goes on. And growth, sometimes significant growth, is possible even in the last 4-6 weeks of the regular season.

Villanova Specifically
Fans may be concerned with the Wildcat's defensive rankings, but progress on offense suggests progress on defense is also possible, indeed even probable. Consider Mouph and Reggie have both just rejoined the team, and both were projected to have large roles -- that has most likely not changed. They will need time to get themselves back into playing shape, and pick up the rhythm of their teammates.

No comments: