Bracket Building 201

In part one of Bracket Building, Kyle Lamb introduces a six-year mock committee to replicate the NCAA Tournament selection process. In part two, the mock committee discloses the results of the first ballots and also find out some interesting numbers at how the AP poll correlates more closely to seeding in the NCAA Tournament than you probably imagined.

As of Wednesday afternoon, the NCAA Tournament selection processes – both authentic and replicated, are well underway. While the 10-person committee full of directors of athletics and conference commissioners assembled in an Indianapolis hotel early Wednesday morning, the "mock" selection committee got a three-day head start.

 

Earlier last week, we introduced you to "Bracket Building 101." The article gave some background and foreground to both the history of our mock selection committee and pending protocol to the upcoming process.

 

That link, found here: http://ohiostate.scout.com/2/623024.html was the ultimate precursor to Bracket Building 201, which will guide you through the actual finer points of the process early in selection week.

 

For nearly six weeks our mock committee has met via online instant messaging preparing for the Super Bowl of aspiring bracketologists everywhere. Scattered around the entire country, our 10-person conglomerate of journalists, journalists-in-training, professionals and college basketball junkies met Sunday evening to officially kick off selection-week 2007.

 

If you don't recall the procedures for this first meeting, allow me to give a refresher course.

 

Due Sunday evening from each of the 10 committee members were two ballots. The first contained up to 34 teams that each individual felt was deserving of an at-large bid based on their successful play to date. Any teams that had already secured an automatic bid into the NCAA Tournament were excluded from the vote.

 

The second ballot contained any other teams that the committee members felt were deserving of at-large consideration. Since there is no set number on how many teams to include on the second ballot, we have loosely defined this ballot as listing any team you could reasonably justify tournament inclusion pending a possible successful conference tournament run.

 

As instructed, members submitted their ballots to a website set up to host our custom ballots, where the chairman could retrieve the results and download them to a spreadsheet, updating the committee workbook in Microsoft Excel format. These workbooks are set up very similar to the real computer software documents used by the NCAA selection committee.

 

In past seasons, our "mock" committee has usually placed between 24 and 28 teams in the at-large field with the first ballot. This ballot requires a team to earn all but two of the eligible votes to gain initial entry in the field as an at-large bid. In other words, if UCLA or Ohio State is represented on the committee, which is accurate on both accounts this season as directors of athletics Dan Guerrero and Gene Smith are both on the NCAA selection committee, then either team would require seven of nine votes for being included on the first ballot.

 

Naturally, that would not be a problem for either team – expected to get No. 1-seeds in the tournament this season.

 

This season, it was smoother sailing the first time around. After all 10 ballots had been received a total of 31 teams received the necessary votes to be added into the field of the mock selection committee after the first step. Here are the results of that first ballot:

 

(Note: all Big East teams, Marquette, Oregon, Texas Tech, Ohio State, Wisconsin and Kentucky all had representatives on the committee.)

 

Bolded teams gained the necessary votes on the first ballot.

 

Air Force

4

Arizona

10

Boston College

10

Butler

10

BYU

7

Drexel

1

Duke

10

Florida

10

Georgetown

9

Georgia Tech

4

Indiana

9

Kansas

10

Kentucky

9

Louisville

9

Marquette

7

Maryland

10

Memphis

10

Michigan State

4

Missouri State

1

Nevada

10

North Carolina

10

Notre Dame

8

Ohio State

9

Oregon

8

Pittsburgh

9

Southern Cal

8

Southern Illinois

9

Stanford

3

Syracuse

3

Tennessee

10

Texas

10

Texas A&M

10

Texas Tech

1

UCLA

10

UNLV

9

Vanderbilt

8

Villanova

5

Virginia

10

Virginia Tech

10

Washington State

10

Wisconsin

10

 

According to the rules, any teams not gaining the necessary votes for the first ballot would automatically be placed on the committee's nomination board for further consideration with at least two votes on this ballot. This means Air Force, BYU, Georgia Tech, Michigan State, Stanford, Syracuse and Villanova were placed on the nomination board regardless of the second ballot.

 

The second ballot, which formed the nomination board, yielded the following results (with qualifying teams in bold):

 

Air Force

6

Akron

1

Alabama

10

Appalachian St.

3

Arkansas

3

Bradley

6

Bucknell

1

BYU

2

Clemson

8

Coll. Of Charleston

1

DePaul

6

Drexel

8

Florida State

6

Georgia

6

Georgia Tech

5

Gonzaga

9

Hofstra

2

Holy Cross

1

Illinois

9

Iowa

4

Kansas State

9

LSU

1

Marquette

1

Massachusetts

9

Michigan

8

Michigan State

6

Mississippi

5

Mississippi State

2

Missouri

1

Missouri State

9

New Mexico State

3

North Carolina St.

1

Oklahoma

1

Oklahoma State

7

Old Dominion

7

Providence

2

Purdue

9

San Diego State

7

Southern Cal

3

Stanford

6

Syracuse

6

Texas Tech

7

Utah State

1

VCU

10

Villanova

4

Washington

4

West Virginia

9

Wichita State

1

Wright State

2

Xavier

8

 

A total of 38 teams were under consideration in addition to the previous 31 after this first critical step. Teams required at least two votes on this or the other ballot to be considered. In other words, Akron, Bucknell, Charleston, Holy Cross, LSU, Missouri, North Carolina State, Oklahoma, Utah State and Wichita State were all dropped from the list.

 

Upon completion of this first step, the mock committee decided to fill our at-large field for the time being. It was determined there were enough teams out there we felt comfortable including even before the beginning of a majority of conference tournaments this week.

 

The next step by rule was for each committee member to list eight (8) teams from the nomination board they felt should be considered for an at-large bid. Any team that gained all but two of the eligible votes would automatically be added to the at-large board with the other 31 teams.

 

Here are those results:

 

Air Force

 

77.80%

7

Alabama

 

11.10%

1

BYU

 

88.90%

8

Drexel

 

11.10%

1

Florida St

 

11.10%

1

Georgia Tech

 

100%

9

Illinois

 

22.20%

2

Kansas St

 

22.20%

2

Michigan St

 

66.70%

6

Missouri St

 

22.20%

2

Old Dominion

 

22.20%

2

Stanford

 

55.60%

5

Syracuse

 

55.60%

5

Texas Tech

 

44.44%

4

Villanova

 

77.80%

7

West Virginia

 

11.10%

1

Xavier

 

55.60%

5

 

This vote dictated, based on the "all but two of the eligible votes policy," that Air Force, BYU, Georgia Tech and Villanova would be added to the at-large board. Since there were already 31 teams, this meant we would have one extra team that would gain a spot once another automatic bid was earned by a team already in the field.

 

Remember, as teams earn automatic bids, they are replaced from the at-large field. Since UCLA was already on the at-large board, should the Bruins win the PAC 10 conference tournament, they would be removed from the at-large board – opening another bid for a team on the nomination board.

 

After this step, the at-large board for the mock committee looks like this:

 

Air Force

Arizona

Boston College

Butler

BYU

Duke

Florida

Georgetown

Georgia Tech

Indiana

Kansas

Kentucky

Louisville

Marquette

Maryland

Memphis

Nevada

North Carolina

Notre Dame

Ohio State

Oregon

Pittsburgh

Southern California

Southern Illinois

Tennessee

Texas

Texas A&M

UCLA

UNLV

Vanderbilt

Villanova

Virginia

Virginia Tech

Washington State

Wisconsin

 

The top four vote getters from the last ballot will be held over for the next ballot, where committee members will select four more from the nomination board. The top four will join those four for a "cross-country" ballot, which will rank them 1-8 with one being the best and eight being the worst. The four teams that will be held over based on the last ballot are Michigan State, Stanford, Syracuse and Xavier.

 

Next, the mock committee elected to eliminate some teams off the nomination board that we felt were wasting our time for at-large bids. A vote to eliminate teams requires the same "all but two" principal of the eligible votes.

 

The committee voted on probably 10 teams, but the teams that were officially eliminated were Hofstra, Mississippi State, New Mexico State, Providence and Wright State. Tuesday evening, Wright State earned a bid anyhow by virtue of their Horizon Conference Championship against Butler. In Tuesday's meeting, the committee also removed Bradley and Oklahoma State from consideration.

 

When it was all said and done, the mock committee currently is left with 25 teams under consideration on the nomination board in addition to the 35 sitting peacefully in our field as at-large qualifiers.

 

Belmont, Winthrop, VCU, Pennsylvania, Niagara, Oral Roberts, Creighton, Eastern Kentucky, Davidson, North Texas and Gonzaga had already earned automatic bids as of Wednesday morning. In addition to the list of 35 teams above, here are the 25 still under consideration according to the mock committee:

 

Alabama

Appalachian St

Arkansas

Clemson

DePaul

Drexel

Florida St

Georgia

Illinois

Iowa

Kansas St

Massachusetts

Michigan

Michigan St

Mississippi

Missouri St

Old Dominion

Purdue

San Diego St

Stanford

Syracuse

Texas Tech

Washington

West Virginia

Xavier

 

As of now, in addition to the one extra spot required to open up on the at-large board for the extra team that's been voted in by the mock committee, it appears a maximum of eight more spots could open up. This means that at very most, only eight (8) of these 25 teams can hope for an at-large bid into the NCAA Tournament – and that's assuming there are no upset winners in conference tournaments that would not have otherwise earned a bid.

 

That's where the field is currently standing. Of course, the authentic selection committee just began Wednesday morning, so while they're a bit further behind, they have all weekend to convene and catch up.

 

Our next meeting will be Thursday evening. It's unlikely we'll make any more votes until Friday, but bubble teams will be closely examined at that time in anticipation of the pressure-filled weekend.

 

New Horizons

 

Some of the most fun by the (mock) selection committee every season is handling the intriguing cases. Teams like Butler are usually tough nuts to crack because of the lack of competition in their own conference.

 

This season, the Bulldogs took college basketball by storm with victories against Indiana, Notre Dame, Purdue and Gonzaga and Tennessee both on a neutral court. Butler, however, has lost four of it's last 10 games including a pair of defeats against Wright State, a loss in the bracket busters to Southern Illinois and a regular season loss to Loyola.

 

With the late-season collapse by Butler, prognosticators are predicting anywhere from a 5-seed to 10-seed in the NCAA Tournament, which will be announced Sunday.

"At this point I see them as no better than an 8 seed," said mock committee member Galen Clavio, an assistant instructor at Indiana University and six-year member of the annual group. "They've done nothing worthwhile since mid-December, and they didn't win anything in a relatively weak Horizon League."

 

But "not so fast my friend," the authentic group might say.

 

Horizon Commisioner Jonathan LeCrone sits on this year's NCAA selection committee. Butler head coach Todd Lickliter serves as the Horizon representative for the regional advisory committee, a panel of 31 coaches from each league to report to the committee regarding teams in that conference.

 

Although steps are taken to remove potential bias, for instance, a commisioner or AD having to leave the room when his or her team (or teams) are being discussed, the mock committee still believes based on past results there is some degree of nepotism in committee members selecting and seeding teams.

 

Statistically speaking, that charge isn't completely out of line given the history of suspect selections. Last season, although George Mason justified their inclusion in the tournament field with a magical Final Four run, many questioned how they could have been picked over Hofstra – who finished with a better record and beat George Mason both times they played. For good measure, George Mason's leading scorer Tony Skinn was suspended and would have missed the first game of the NCAA Tournament, which is criteria the committee has always claimed to take into account.

 

Also last season, Air Force and Utah State made the tournament as at-large selections despite low strength of schedule numbers (158 and 102 respectively), The best win for Air Force was a home victory against No. 56 San Diego State. Meanwhile, Utah State had only six wins against the RPI top 100.

 

Both had representation on the committee as did George Mason, whose director of athletics Tom O'Connor served as a 2006 selection committee member.

 

In Wednesday's Indianapolis Star, NCAA Vice President Greg Shaheen, who also helps with administrative matters of the NCAA selection committee, defended that selection and explained the committee's rationale. He unknowingly contradicted the perception that the committee was numbers driven, objective in analyzing numbers, data, and team profiles.

 

"Ultimately, (the committee) backed off the numbers and talked about basketball quality, in their minds," Shaheen told the paper about the George Mason inclusion.

 

But perhaps the RPI, how you fare in your last 10 games, road records and the like are only a small piece of a very large, subjective puzzle.

 

"People succumb to this notion of the 'tyranny of numbers,'" added Princeton AD Gary Walters, who is serving as this year's chairperson. "You have to be careful about that. My wife is a CPA, and she warned me early in my career, when I was in the investment business, that if you torture the numbers long enough, you can get them to confess to anything."

 

Crunching (or compacting) Numbers

 

George Mason is far from the most controversial selection as an at-large bid into the NCAA Tournament. That honor easily goes to New Mexico in 1999.

 

New Mexico made the tournament as an at-large with an all-time high RPI of 74 at that time. Their selection, which was controversial based on their RPI, strength of schedule rated No. 157 and also their loss to No. 215-rated Hawaii, made it an intriguing choice.

 

Making the choice noteworthy was the presence of New Mexico AD Rudy Davalos, who served on the 1999 committee.

 

Clavio, like many other people nationwide, believes the "subjective" criteria have often outweighed the material numbers and procedure.

 

"The biggest myth is this notion of ‘serendipity' – that the committee isn't aware of the potential matchups they create in the seeding and placement process," Clavio explained. "If you're knowledgeable enough about college basketball to be able to rank 65 teams in order, then you certainly know enough to realize what combinations of teams will be more compelling than others."

 

That's certainly understandable given CBS Sports' stake in their $545 million-per-season contract with the NCAA to broadcast the tournament. But despite the claims from the NCAA that the committee does not consult CBS during the selection and the field is not put together with matchups in mind, it's hard not to take that into account when bracket rules permit and opportunities exist.

 

Perhaps a bigger myth – think the AP poll doesn't matter in college basketball? Directly or indirectly, that might not be accurate.

 

After growing curiosity over sneaking suspicions, I researched the final regular season AP polls over the last 10 seasons and compared them to seeding in the corresponding NCAA Tournaments.

 

The results were staggering.

 

The NCAA Tournament is put together after selection and subsequent 1-65 ranking of all the teams. That process, which is called the S-curve, will be explained later in the week of "Bracket Building," but suffice to say, if you're ranked 1-4, you will be a No. 1 seed; 5-8 a No. 2 seed; 9-12 a No. 3 and so on and so forth.

 

There is some leeway for moving teams up or down one seed based on necessity to obey bracket rules, but generally, this is rarely needed.

 

A study of the last 10 seasons shows that if the AP top 25 poll seeded those 25 teams, just as the s-curve does for all 65 teams, the brackets potentially would look eerily similar.

 

Of 250 teams ranked in the final regular season poll in 10 years, 214 of those teams were seeded in that season's NCAA Tournament within one seed of their corresponding AP rankings. That's 86 percent.

 

Of those 36 instances where a team was more than two seeds off their corresponding AP ranking, treating the AP as if it were the S-curve ranking of the NCAA field, 19 of them were teams from mid-major or non-BCS conferences that had questionable schedule strengths, lower RPI rankings or simply had questions regarding how good they really were.

 

What's more is that of 250 teams, 112 of them (45 percent) were seeded in the NCAA Tournament exactly at the same seed as they would have been if the AP ranking dictated the top 25 entirely.

 

What does this suggest?

 

It means the NCAA committee likely bases their rankings from the associated press poll and makes changes here or there to customize their thoughts on the numbers, subjectivity or the criteria of the day.

 

In the last 10 years, you can usually count on the No. 1 and No. 2 seeds in the NCAA Tournament coming from the top-8 of the AP poll. In 68 of 80 occasions over the last 10 years, the top two lines of the NCAA Tournament bracket has included teams ranked in the top eight of the AP. Thirty four of 40 teams that finished the regular season in the AP top four ended up a No. 1-seed in the NCAA Tournament.

 

That means by average, three of the AP top four will be a No. 1-seed in the NCAA Tournament and the fourth No. 1 seed will come from the top eight. This year's AP poll currently lists Ohio State, Kansas, Wisconsin and UCLA as the top four seeds. Memphis, Florida, Texas A&M and North Carolina round out the rest of the top eight.

 

Bracket Building 301 will be back Friday with more from the mock selection committee. Next up for the committee: examining the remaining 25 teams on the nomination board with a fine tooth comb, preparing for another 6-8 spots to open up on the at-large board this weekend and debate seeding – which will likely begin for the mock committee Friday evening.

 

 

 

 

 

 

 


Buckeye Sports Top Stories