|
Contributed by Peter Tiernan, ESPN.com Bracketologist
|
We examined which single attributes led to the greatest overachievement in the NCAA Tournament and we discovered that factors like victory margin, coaching experience and "star presence" were the top indicators of seed-defying runs. Of course, most tourney fanatics will consider more than one attribute in making their bracket picks. With that in mind, we look at which pair of attributes lead to the best performance against seed expectations (PASE).
When you evaluate tourney teams based on a pair of factors, two countervailing things happen:
To illustrate how much more valuable PASE attribute pairs are than single attributes, take a look the PASE values of the top five attribute pairings compared to the top five single pairings:
Teams with a victory margin of more than 15 points have a PASE value of +.424, suggesting that more than 40 percent of these teams surpass seed-projected win totals by one game. The second best single-attribute PASE (a coach with 15 or more tourney appearances) yields a +.213 PASE. Where would that fall on the list of top attribute pairings? How does 35th place sound? As the red line demonstrates, the top five attribute pairings have a markedly higher PASE value than their single-attribute counterparts (though there's no denying that victory margin remains a powerful factor).
Here's the rub, though: Whereas 102 teams have satisfied the victory margin attribute -- about five per tourney -- only 63 have met the top dual-attribute conditions (more than 60 percent of points from the frontcourt and one preseason All-American). That works out to fewer than three teams fulfilling the pairing conditions per tourney. The lesson? As we examine the top 10 overachievement attribute pairings, pay special attention to the number of teams that satisfy each condition. This will give you a clear indication of how likely you are to find teams possessing each combination of factors.
If you were looking for 2006 teams that satisfied the "frontcourt strength/star power" pairing, you'd find three candidates: Connecticut, West Virginia and Gonzaga. The Mountaineers were the highest overachiever of the three, winning two games as a No. 6 seed. The Zags barely overachieved with a Sweet 16 run as a No. 3 seed. And the Huskies slightly underachieved, winning only three games, where the average top seed gets 3.32 wins per tourney. Their collective PASE was +.205, well under the +.509 PASE, but still solid overachievement.
Two other attribute pairings yield a PASE above +.400. The first one -- scoring more than 77 points and beating teams by more than 15 points a game -- shouldn't come as a surprise. Both scoring offense and victory margin were identified as key single-attribute overachievement indicators. The second pairing -- a winning percentage above .850 with "strength of schedule" better than 50th place -- is surprising because the two factors ranked seventh and 30th respectively on the list of 56 individual attributes studied. When combined, however, the attributes yield a +.404 PASE.
Unfortunately, you're not likely to find more than two or three teams that satisfy these conditions each tourney. In 2006, three top seeds qualified (UConn, Duke and Memphis) -- and they all underachieved. You'll usually find more than twice as many teams with the requisite offensive punch and margin qualifications (93) -- plus, their likelihood of overachieving is a notch better (+.426 PASE). Surprisingly, only one team, UConn again, possessed the "scoring/margin" qualifications, but Florida came mighty close, averaging 79.3 points a game, but only beating opponents by 14.5.
If it's quantity you're after, look at teams with solid tourney experience and a season-ending loss before the dance. There are 135 teams with more than four straight bids who came into the tourney with a one-game losing "streak" (most likely in a conference tournament). That works out to more than six teams per dance and those squads have a solid +.374 PASE, good enough for fifth place on the PASE attribute pairings list. Last year, seven teams met these requirements. Their performance was mixed. UConn, Illinois, Pitt and Michigan State underachieved. Texas, Arizona and Kentucky overachieved.
The next most prevalent attribute pairing on the top 10 list combines victory margin and star power. There are 99 squads -- 4.5 per tourney -- who beat their opponents by more than 12 points a game and boast a preseason All-American. These teams win 31 more games than their seed-projected total of 270 for a solid +.311 PASE. The 2006 teams that fell into this category were UConn, Duke, Memphis -- all underachievers -- and Washington, an overachiever with two wins as a five seed.
What trio of attributes leads to the highest overachievement?
The top trio of attributes that lead to tourney overachievement are frontcourt dominance, star power and coaching experience. The 22 teams that get more than 60 percent of their points from their frontcourt, have a preseason All-American and a coach with more than 10 tourney bids win 20 more games than the 56 they're expected to win. That works out to a near-one-game-per-tourney PASE of +.931. Last year, only UConn possessed these three attributes and they failed to overachieve. The year before, however, tourney-champ North Carolina, surprise semifinalist No. 4 seed Louisville and No. 2 seed UConn possessed the top trio of attributes. Taken together, the four qualifying teams from 2005 and 2006 had a PASE of +.855 -- just under the overall +.931 value.
As it turns out, UConn is the only 2006 team to satisfy the first four attribute trifectas on the list above. Both UConn and No. 5 seed Washington met the requirements of the fifth-place trio of attributes (star power, victory margin and one-game losing streak) -- and Washington's version of Huskies were tourney overachievers with two wins.
Texas was the only other team from 2006 to satisfy any of the top 10 attribute trios. The Longhorns met the eighth-place requirements of frontcourt dominance, victory margin and one-game losing streak. With three wins as a No. 2 seed, they were a PASE overachiever. So altogether, the three teams that possessed one of the trios of attributes on the top 10 list were UConn, Washington and Texas. Two of these three overachieved, and their overall PASE was +.375.
Looking ahead to the 2007 tourney
The thorniest issue that tourney pool players will face when they fill out their brackets in March is defining what a "preseason All-American" really is. It used to be that the AP publicized the top 10-15 players on their preseason list. In recent years, they've taken to announcing only the top five and one "sixth man." This year, Tyler Hansbrough (North Carolina), Joakim Noah (Florida), Glen Davis (LSU), Ronald Steele (Alabama) and Brandon Rush (Kansas) were the top five vote getters, with Aaron Afflalo (UCLA) netting the sixth-highest vote total.
With the "star power" analysis identifying 10-15 players in nearly two-thirds of the 64-team era, the last few years have been short-changed for this requirement. This year, we'll be moving to a more reliable indicator of star power: the United States Basketball Writers Association All-American list, which is typically announced at least a week before the tourney. The advantage of the USBWA list is twofold:
Adding another four players to the "All-American" qualification could give a host of other teams the opportunity to satisfy the key PASE attribute pairings and trios. Pitt, Georgetown, Ohio State, Boston College, Arizona, Texas and Nevada are just a few of the teams with star players who could challenge the usual suspects for dominance in this year's tourney.
Freelance writer Pete Tiernan has been studying the NCAA Tournament for 17 years. E-mail him here. Or go to www.bracketscience.com and check out the BracketMaster, which lets you do your own bracket analysis. | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
