Bracketology Mythology: Comparing the Predictions for the 2012 NCAA Basketball Brackets

The term “bracketology” has become a popular term in the sports lexicon, especially during the months of February and March.  According to the most trustworthy source of knowledge in the world, “bracketology” means

the process of predicting the field of the [Men’s] NCAA Basketball Tournament (wikipedia).

It is a term that was popularized by Joe Lunardi, who currently works for ESPN, but a term that is used by all.  In fact, bracketology itself has practically become a sport of its own, with several websites touting its expertise.

Well, how accurate are these predictions?  Glad that I asked.  That is what I decided to look at with this entry.  While there are some websites that breakdown the success/failure rate of these bracketologist, I am more concerned with the top sports media websites with regards to men’s college basketball — ESPN, Yahoo! Sports, and CBS Sports.  I want to look at these in particular because most people likely turn to these sites to get their projections.  I also include the website Bracketville, which according to The Bracket Project is the most accurate projector of brackets among 44 “veterans” [Yahoo! Sports’ Mike Huguenin ranks 22nd; CBS Sports’ Jerry Palm is 26th; Lunardi for ESPN is tied for 27th].

Before I get started, it is important to understand what we are looking at here.  First, I looked at the difference between each bracketologist’s projection and the actual seeding.  With this, I could figure out how accurate each site proved to be.  In order to account for teams missed (like Iona), the bracketologist is given a “predicted” seeding of 17 [i.e. not in the tournament].  Second, I took a total count of accurate seedings.  Third, using a scoring system established by PHSports, a score is giving for each bracketologist that also includes total number of teams in the field, accurate seedings, and seedings with one line.  Finally, for ESPN, CBS Sports, and Bracketville, I also looked at accuracy with seedings AND regions [Yahoo! Sports does not slot teams into regions].

So, in terms of accurately projecting seeds, Jerry Palm was the most accurate by getting 36 seeds correct.  However, the other three bracketologists landed 35 seeds.  Therefore, there is no big difference between the four guessers.  On average, all tended to underseed teams, although this is somewhat tricky.  Obviously missing out on a team is going to bring down the number and grossly skew towards underseeding.  Because all missed on Iona, all are knocked down a level.  But, Huguenin, Palm and Bracketville all missed on another team.

  • Bracketville: -8
  • CBS Sports (Palm): -8
  • ESPN (Lunardi): -3
  • Yahoo! Sports (Huguenin): -8

Now, if we take out the teams missed completely, most projections naturally level off, with Palm overseeding by one and Bracketville underseeding by 2.  In terms of teams, the biggest overseeded teams are BYU (Huguenin and Palm by three) and Creighton (Palm by three) and the biggest underseeded team was Southern Miss (Huguenin by three).  In terms of total over/underseeding (adding up the seeding differences between the four bracketologist), Harvard can be added to BYU and Creighton as teams overseeded (all overseeded Harvard by two spots).

So, what about the bracketology score?  While all nailed roughly the same number of seeds correctly, there is a slight difference when it comes to within one seed line.

  • Bracketville: 351
  • CBS Sports (Palm): 352
  • ESPN (Lunardi): 358
  • Yahoo! Sports (Huguenin): 351

In this case, Lunardi’s extra points for missing out only on Iona helps him here.  Palm gets a leg up on the other two for getting one more seeding correct.

Finally, let’s look at the three that broke down the regions.  Here are the top four seeds in each region:

EAST

SEED

Lunardi

Palm

Bracketville

ACTUAL

1

Syracuse

Syracuse

Syracuse

Syracuse

2

Duke

Duke

Duke

Ohio State

3

Michigan

Michigan

Michigan

Florida State

4

Indiana

Wisconsin

Indiana

Wisconsin

SOUTH

SEED

Lunardi

Palm

Bracketville

ACTUAL

1

North Carolina

North Carolina

North Carolina

Kentucky

2

Ohio State

Ohio State

Ohio State

Duke

3

Baylor

Baylor

Baylor

Baylor

4

Wisconsin

Indiana

Georgetown

Indiana

MIDWEST

SEED

Lunardi

Palm

Bracketville

ACTUAL

1

Kentucky

Kentucky

Kentucky

North Carolina

2

Missouri

Missouri

Missouri

Kansas

3

Georgetown

Louisville

Louisville

Georgetown

4

Louisville

Florida State

Wisconsin

Michigan

WEST

SEED

Lunardi

Palm

Bracketville

ACTUAL

1

Michigan State

Michigan State

Michigan State

Michigan State

2

Kansas

Kansas

Kansas

Missouri

3

Marquette

Marquette

Marquette

Marquette

4

Florida State

Murray State

Florida State

Louisville

All missed on the location for Kentucky and North Carolina, as well as all of the number two seeds.  Although, they did at least get those seeds correct.  In fact, in terms of the top four seeds in each region (i.e., top 16 seeds overall), the only mis-seeded teams are Florida State (all projected the Seminoles fourth rather than third) and Michigan (third rather than fourth).  Only one person got Louisville in the correct fourth seed (Lunardi), while only one person nailed Georgetown in the three seed (Lunardi again).  Murray State (Huguenin and Palm) was the only team predicted to be a top four seed but missed, while Indiana (Huguenin) and Georgetown (Palm) were the only teams predicted outside of the top four seeds.

As far as accuracy in regions, Lunardi and Palm both got 16 teams in their correct regionandseed, while Bracketville only got ten teams.  They were the most accurate in the East bracket, with Lunardi getting five correct, Palm with four teams, and Bracketville with three teams.  The fewest accurate projections was Jerry Palm’s Midwest bracket (Vermont as the sixteenth).  In terms of at least getting the teams in the correct region regardless of proper seeding, Palm was the most accurate with 25 teams in the correct region, followed by Lunardi (21) and Bracketville (14).

One final measure of region/seeding accuracy, using the scoring measure for the regions lends itself heavily towards Palm.

  • Bracketville: 78
  • CBS Sports (Palm): 137
  • ESPN (Lunardi): 121

Now, the trickiest part is the Last Four In/First Four Out.  According to the First Four match-ups, the Last Four In are BYU, California, Iona and South Florida.  With regards to this, Bracketville (California and South Florida) and Lunardi (BYU and South Florida) got two right, while Palm only got one (California).  In terms of their predicted Last Four actually making the field in any capacity, Bracketville and Lunardi add North Carolina State, while Palm added Colorado State.  All three missed on Seton Hall, while Palm (Northwestern) and Bracketville (Drexel) missed on another.

As for First Four (or more) Out, according to ESPN.com, the NCAA would have chosen between Drexel, Miami (FL), Mississippi State, Nevada and Oral Roberts had St. Bonaventure not won the Atlantic-10 tournament.  Nevertheless, Lunardi and Palm both missed on Iona while Bracketville did not even list Iona as in consideration.  Palm also missed on NC State while Bracketville missed BYU.

So, in terms of general accuracy, it does not seem that there is much difference between the four bracketologists, at least for the 2012 Tournament.  But, when we look at regional accuracy, Palm and Lunardi are far more accurate than Bracketville.

The last thing I want to look at are the teams; which teams did all sources accurately project and which teams did they all miss?  Well, there were 20 teams that all four bracketologists accurately predicted their seeding:

ALL CORRECT

Team

Rank

Team

Rank

Alabama

9

Mississippi Valley State

16

Baylor

3

Missouri

2

Duke

2

North Carolina

1

Kansas

2

Ohio State

2

Kentucky

1

Saint Louis

9

Lamar

16

Syracuse

1

Lehigh

15

Vermont

16

Loyola (MD)

15

West Virginia

10

Marquette

3

Western Kentucky

16

Michigan State

1

Wisconsin

4

And then, there were 13 teams that none of the four got right (n is the average mis-seeding):

ALL WRONG

Team

n

Team

n

BYU

1.25

Memphis

1.25

Creighton

2

Michigan

1

Florida State

-1

Montana

-1

Gonzaga

-1

Norfolk State

-1

Harvard

2

Southern Miss

-1.75

Iona

-3

St. Bonaventure

1

LIU-Brooklyn

1

As mentioned above, Iona surprised just about everyone with their inclusion and was the only team that all four bracketologists left off of their projections.  The only other teams left off were BYU (Bracketville), California (Huguenin), and North Carolina State (Palm).

And finally, here are the teams that only one of the Bracketologists got correct:

  • Bracketville [3]: Colorado State (11), Murray State (6), UNLV (6)
  • CBS Sports (Palm) [4]: California (12), Cincinnati (6), New Mexico State (13),VCU (12)
  • ESPN (Lunardi) [4]: Georgetown (3), Louisville (4), Purdue (10), Texas (11)
  • Yahoo! Sports (Huguenin) [7]: Colorado (11), Davidson (13), North Carolina State (11), Notre Dame (7), San Diego State (6), Temple (5), Virgina (10)

Okay, that is all.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s