- Mar 19, 2006
- 38,712
- 599
- 113
- 75
Bracket Science: Searching for likely upset victims, victors
<TABLE class=storyHeader border=0 cellSpacing=0 cellPadding=0 width="100%"><TBODY><TR><TD class=storyInfo vAlign=top>March 15, 2010
By Peter Tiernan
Special to CBSSports.com
</TD><TD class=storySponsor><TABLE cellSpacing=0 cellPadding=0 width=0><TBODY><TR><TD><!-- 400x60 --><SCRIPT type=text/javascript>var cbsiad224_100 = { 'SP' : "224", 'POS' : "100"};if (location.search.substring(1).indexOf('DCLK')>-1) document.write('<input type="text" value="cbsiad224_100" style="width:400px">
');cbsiGetAd(cbsiad224_100); </SCRIPT><SCRIPT type=text/javascript src="http://mads.cbssports.com/mac-ad?CLIENT:ID=SJS&celt=js&x-cb=40170717&IREFER_HOST=cbssports.com&SITE=175&BRAND=59&NCAT=22054%3A22164%3A22573%3A23598&NODE=23598&PTYPE=6858&CID=13063296&DVAR_SESSION=a&DVAR_SUBSESSION=2&cookiesOn=1&DVAR_EXCLUDE=ncaabb&DVAR_USER=anon&DVAR_PRINTPAGE=1&STAGING=0&SP=224&POS=100&COOKIE%3AANON_ID=Cg8IMEuet3IHAAAAbAs"></SCRIPT><!-- default ad -->
<NOSCRIPT></NOSCRIPT> </TD></TR></TBODY></TABLE></TD></TR></TBODY></TABLE><TABLE border=0 cellSpacing=0 cellPadding=0 width="100%"><TBODY><TR vAlign=top><TD>[FONT=Arial, Helvetica]
[FONT=Arial, Helvetica]<!-- T13063296 --><!-- Sesame Modified: 03/15/2010 17:47:04 --><!-- sversion: 8 $Updated: fagan$ -->We've been hearing for weeks now that this year's tourney will be crazier and more unpredictable than the relatively chalky dances of the past three years. Possession-based statistics confirm that the higher seeds, from one all the way to seven, are weaker than past squads, while the lower seeds, from eight to 13, are stronger. [/FONT]
If you're going to win your pool this year, you're going to have to pick some upsets. The question is: which ones? When the bracket is staring you in the face, it's hard to pull the trigger on a big giant killing ... simply because you know the damage it can do to your chances if you're wrong. I found it ironic that the same experts who've been saying all along that 2010 is going to be a mad dance stuck with one and two seeds in their Final Four choices.
Last week, we ran an article called "Anatomy of an Upset" that identified the historical conditions leading to upsets in the first and second rounds of the dance. I wrote then that I would report which teams in the 2010 field possess the qualities of upset victors -- and which teams showed warning signs of being victims. Here's what the numbers say for the first two rounds:
First-round victors and victims
If you read the earlier piece you know that I argued about the folly of picking any upset lower than a 4-13. One seeds are a lock against 16 seeds --100-0 over the 25 years of the 64-team era; two seeds are nearly as strong against 15 seeds, at 96-4; and three seeds are a dominant 85-15 against 14 seeds. The last top-three seed to get upset in the tourney was Iowa in 2006, when the Hawkeyes lost to No. 14 Northwestern State. I don't see any of the three seeds getting shocked this year (if you forced me to pick one, I would say New Mexico), so let's concentrate on the 6-11, 5-12 and 4-13 games.
In the 6-11 matchups, only one squad has the Cinderella qualities we cited in Anatomy of an Upset. The Washington Huskies average more than 73 points per game and have an average scoring margin greater than seven points.
[SIZE=+1]? In the 5-12 matchups, two teams have the right stuff to spring a surprise. Both Cornell and Utah State went to the dance last year and get at least 51 percent of their points from the frontcourt. [/SIZE]
[SIZE=+1]? In the 4-13 matchups, watch out for Murray State. The Racers have an experienced tourney coach, the requisite scoring margin, and the backcourt/frontcourt scoring balance to slay a giant.
Those are the Cinderella teams with the right numbers to pull off an upset. Are there any favored teams that show victim-like warning signs? Surprisingly, none of the six seeds have historical vulnerabilities. Among five seeds, however, Temple is offensively challenged enough to qualify as a victim -- and, lo and behold, the Owls play Cornell in the first round. A similar victor/victim alignment can be seen in one of the 4-13 matchups. Vanderbilt has the requisite lack of experience to get toppled in Round 1. Their opponent? You guessed it: those pesky Murray State Racers. Maryland also meets the upset criteria, but the Terps are a longer shot to be slain, since their opponent, Houston, doesn't have the numbers to pull off a stunner.
Second-round victors and victims
In "Anatomy," I pointed out that there are really only four matchups worth focusing on for a second-round Cinderella -- the 1-8, 2-7, 2-10 and 3-11 tilts. In the possible 1-8 and 2-10 matchups, none of the lower seeds have the qualities of an upset winner. In the 2-7 matchups, both Oklahoma State and Clemson meet the conditions of coming from a Big Six conference with a scoring margin better than six points per game and solid pre-tourney momentum. Beware of Clemson, though: Oliver Purnell is the all-time underachieving coach. He fails to meet seed-expected win totals by more than three-quarters of a game per dance.
As for the 3-11 matchups, both Minnesota and Washington have what it takes to pull off a second-round upset. They're from Big Six conferences, have tourney-tested coaches, decent momentum and good scoring balance between their guards and big men
Top-three seed underachievers
At the tail end of "Anatomy," I listed the signs of one-, two- and three-seed underachievers. This doesn't necessarily mean that these squads have been upset in the dance; just that they've failed to live up to seed expectations.
Top-three seeds have won 3.44 games per tourney since 1985, so failing to reach the Final Four qualifies as underachievement for a top seed. Are there any No. 1s that show warning signs of an early exit? In fact, there are two. Syracuse makes the list because they're too reliant on frontcourt scoring. Duke is also in jeopardy of underachieving -- and, if it happened, that would make it the sixth straight year of failing to meet seed expectations. The Blue Devils' issue? Top seeds that shoot lower than .465 from the field tend to underperform, and Duke's field-goal percentage is .438. Yes, they do shoot a high percentage of 3s and get a lot of offensive boards to extend possessions -- but the numbers are the numbers.
As for two seeds, cast a cold eye on Villanova. Second-seeded squads coming into the tourney with five or fewer wins in their 10 pre-tourney games tend to underperform. (Two seeds have won 2.43 games per dance, so underachievement means failing to make the Elite Eight.) So do teams entering the dance with a losing streak of two or more games. The Wildcats meet both criteria. Surprisingly, Big East champ West Virginia is the other two seed that has the earmarks of an underachiever. The Mountaineers have a field-goal percentage lower than .450, and that has historically been a problem.
That brings us to the three seeds. Amazingly, each of the teams has at least one sign of an underachiever. (Three seeds have won 1.86 games per tourney, which means underachievers fail to reach the Sweet 16.) Georgetown, Baylor and New Mexico didn't participate in last year's tourney, and Pitt has a road/neutral record of .500 or worse (granted, it's exactly .500 at 7-7, but, statistically speaking, my hands are tied).
Those are the teams that history says are most likely to be upset victors and victims and the higher seeds that could suffer an early tourney exit. Of course, these are just historical correlations. The statistics aren't necessarily the cause for a team springing an upset or getting shocked. One thing is always true about upsets -- they're a surprise. I don't see that changing this year. My guess is that we'll have more upsets in 2010 than the historical average of eight per dance ... and many of the Cinderellas will be teams no one expected.
Peter Tiernan has been using stats to analyze March Madness for 20 years. His insights into the NCAA basketball tournament can help you build a better bracket. E-mail him at bracketscience@comcast.net or visit bracketscience.com.
[/SIZE]
</TD></TR></TBODY></TABLE>
<TABLE class=storyHeader border=0 cellSpacing=0 cellPadding=0 width="100%"><TBODY><TR><TD class=storyInfo vAlign=top>March 15, 2010
By Peter Tiernan
Special to CBSSports.com
</TD><TD class=storySponsor><TABLE cellSpacing=0 cellPadding=0 width=0><TBODY><TR><TD><!-- 400x60 --><SCRIPT type=text/javascript>var cbsiad224_100 = { 'SP' : "224", 'POS' : "100"};if (location.search.substring(1).indexOf('DCLK')>-1) document.write('<input type="text" value="cbsiad224_100" style="width:400px">
');cbsiGetAd(cbsiad224_100); </SCRIPT><SCRIPT type=text/javascript src="http://mads.cbssports.com/mac-ad?CLIENT:ID=SJS&celt=js&x-cb=40170717&IREFER_HOST=cbssports.com&SITE=175&BRAND=59&NCAT=22054%3A22164%3A22573%3A23598&NODE=23598&PTYPE=6858&CID=13063296&DVAR_SESSION=a&DVAR_SUBSESSION=2&cookiesOn=1&DVAR_EXCLUDE=ncaabb&DVAR_USER=anon&DVAR_PRINTPAGE=1&STAGING=0&SP=224&POS=100&COOKIE%3AANON_ID=Cg8IMEuet3IHAAAAbAs"></SCRIPT><!-- default ad -->
<SCRIPT type=text/javascript src="http://sports.cbsimg.net/yui/latest/dragdrop/dragdrop-min.js"></SCRIPT><SCRIPT type=text/javascript src="http://sports.cbsimg.net/yui/latest/container/container-min.js"></SCRIPT><SCRIPT type=text/javascript src="http://sports.cbsimg.net/yui/latest/connection/connection-min.js"></SCRIPT><SCRIPT type=text/javascript src="http://sports.cbsimg.net/yui/latest/json/json-min.js"></SCRIPT><LINK rel=stylesheet type=text/css href="http://sports.cbsimg.net/css/email/send-modal.css"><SCRIPT type=text/javascript src="http://sports.cbsimg.net/js/CBSi/app/Email/SendPanel-min-v0001.js"></SCRIPT>
[/FONT][FONT=Arial, Helvetica]<!-- T13063296 --><!-- Sesame Modified: 03/15/2010 17:47:04 --><!-- sversion: 8 $Updated: fagan$ -->We've been hearing for weeks now that this year's tourney will be crazier and more unpredictable than the relatively chalky dances of the past three years. Possession-based statistics confirm that the higher seeds, from one all the way to seven, are weaker than past squads, while the lower seeds, from eight to 13, are stronger. [/FONT]
If you're going to win your pool this year, you're going to have to pick some upsets. The question is: which ones? When the bracket is staring you in the face, it's hard to pull the trigger on a big giant killing ... simply because you know the damage it can do to your chances if you're wrong. I found it ironic that the same experts who've been saying all along that 2010 is going to be a mad dance stuck with one and two seeds in their Final Four choices.
Last week, we ran an article called "Anatomy of an Upset" that identified the historical conditions leading to upsets in the first and second rounds of the dance. I wrote then that I would report which teams in the 2010 field possess the qualities of upset victors -- and which teams showed warning signs of being victims. Here's what the numbers say for the first two rounds:
First-round victors and victims
If you read the earlier piece you know that I argued about the folly of picking any upset lower than a 4-13. One seeds are a lock against 16 seeds --100-0 over the 25 years of the 64-team era; two seeds are nearly as strong against 15 seeds, at 96-4; and three seeds are a dominant 85-15 against 14 seeds. The last top-three seed to get upset in the tourney was Iowa in 2006, when the Hawkeyes lost to No. 14 Northwestern State. I don't see any of the three seeds getting shocked this year (if you forced me to pick one, I would say New Mexico), so let's concentrate on the 6-11, 5-12 and 4-13 games.
In the 6-11 matchups, only one squad has the Cinderella qualities we cited in Anatomy of an Upset. The Washington Huskies average more than 73 points per game and have an average scoring margin greater than seven points.
[SIZE=+1]? In the 5-12 matchups, two teams have the right stuff to spring a surprise. Both Cornell and Utah State went to the dance last year and get at least 51 percent of their points from the frontcourt. [/SIZE]
[SIZE=+1]? In the 4-13 matchups, watch out for Murray State. The Racers have an experienced tourney coach, the requisite scoring margin, and the backcourt/frontcourt scoring balance to slay a giant.
Those are the Cinderella teams with the right numbers to pull off an upset. Are there any favored teams that show victim-like warning signs? Surprisingly, none of the six seeds have historical vulnerabilities. Among five seeds, however, Temple is offensively challenged enough to qualify as a victim -- and, lo and behold, the Owls play Cornell in the first round. A similar victor/victim alignment can be seen in one of the 4-13 matchups. Vanderbilt has the requisite lack of experience to get toppled in Round 1. Their opponent? You guessed it: those pesky Murray State Racers. Maryland also meets the upset criteria, but the Terps are a longer shot to be slain, since their opponent, Houston, doesn't have the numbers to pull off a stunner.
Second-round victors and victims
In "Anatomy," I pointed out that there are really only four matchups worth focusing on for a second-round Cinderella -- the 1-8, 2-7, 2-10 and 3-11 tilts. In the possible 1-8 and 2-10 matchups, none of the lower seeds have the qualities of an upset winner. In the 2-7 matchups, both Oklahoma State and Clemson meet the conditions of coming from a Big Six conference with a scoring margin better than six points per game and solid pre-tourney momentum. Beware of Clemson, though: Oliver Purnell is the all-time underachieving coach. He fails to meet seed-expected win totals by more than three-quarters of a game per dance.
As for the 3-11 matchups, both Minnesota and Washington have what it takes to pull off a second-round upset. They're from Big Six conferences, have tourney-tested coaches, decent momentum and good scoring balance between their guards and big men
Top-three seed underachievers
At the tail end of "Anatomy," I listed the signs of one-, two- and three-seed underachievers. This doesn't necessarily mean that these squads have been upset in the dance; just that they've failed to live up to seed expectations.
Top-three seeds have won 3.44 games per tourney since 1985, so failing to reach the Final Four qualifies as underachievement for a top seed. Are there any No. 1s that show warning signs of an early exit? In fact, there are two. Syracuse makes the list because they're too reliant on frontcourt scoring. Duke is also in jeopardy of underachieving -- and, if it happened, that would make it the sixth straight year of failing to meet seed expectations. The Blue Devils' issue? Top seeds that shoot lower than .465 from the field tend to underperform, and Duke's field-goal percentage is .438. Yes, they do shoot a high percentage of 3s and get a lot of offensive boards to extend possessions -- but the numbers are the numbers.
As for two seeds, cast a cold eye on Villanova. Second-seeded squads coming into the tourney with five or fewer wins in their 10 pre-tourney games tend to underperform. (Two seeds have won 2.43 games per dance, so underachievement means failing to make the Elite Eight.) So do teams entering the dance with a losing streak of two or more games. The Wildcats meet both criteria. Surprisingly, Big East champ West Virginia is the other two seed that has the earmarks of an underachiever. The Mountaineers have a field-goal percentage lower than .450, and that has historically been a problem.
That brings us to the three seeds. Amazingly, each of the teams has at least one sign of an underachiever. (Three seeds have won 1.86 games per tourney, which means underachievers fail to reach the Sweet 16.) Georgetown, Baylor and New Mexico didn't participate in last year's tourney, and Pitt has a road/neutral record of .500 or worse (granted, it's exactly .500 at 7-7, but, statistically speaking, my hands are tied).
Those are the teams that history says are most likely to be upset victors and victims and the higher seeds that could suffer an early tourney exit. Of course, these are just historical correlations. The statistics aren't necessarily the cause for a team springing an upset or getting shocked. One thing is always true about upsets -- they're a surprise. I don't see that changing this year. My guess is that we'll have more upsets in 2010 than the historical average of eight per dance ... and many of the Cinderellas will be teams no one expected.
Peter Tiernan has been using stats to analyze March Madness for 20 years. His insights into the NCAA basketball tournament can help you build a better bracket. E-mail him at bracketscience@comcast.net or visit bracketscience.com.
[/SIZE]
</TD></TR></TBODY></TABLE>
