Bracket Analysis (03-07-2013)

Director of Research and Analytics, John Ewing, takes a look at what the men's college basketball tournament could look like through 50,000 simulations of Bracketology. Follow John on Twitter @JohnEwing for more notes from his analysis and to let him know what else you may like to see with these projections. 

Bracket Odds, Predictalated Bracket, Bracket Analysis

See: How it Works, a Tournament Summary and Regional Previews: South, West, East and Midwest.

How it Works
The Predictalator uses current rosters and strength-of-schedule and pace-adjusted team and player stats (weighted slightly more toward recent games), to play, one possession at a time, every game 50,000 times before it's actually played. For this analysis, we are tracking how likely a team is to make it to any level of the NCAA tournament. Each tournament is played individually, with the team that wins the game in that instance advancing. See the Bracket Odds and Predictalated Bracket.

Tournament Summary

Gonzaga is ranked #1 in the Associated Press Top 25 poll for the first time in school history (they were ranked #2 last week, also the first time in school history). Gonzaga has the best record in Division I at 29-2. The Bulldogs switched places with Indiana after the Hoosiers lost to Minnesota last week. Gonzaga became the third team from a mid-major conference in the last ten years to be ranked #1; the other two schools were Memphis (2007-08) and Saint Joseph’s (2003-04). The Bulldogs ascension to the top spot comes in a year that has seen five programs ranked number one. Indiana held the top spot for 9 weeks, Duke – 5 weeks, Louisville – 1 week, Michigan -1 week, and now Gonzaga – 1 week.

I bring this up because while everyone, including myself, enjoys a great underdog story (this has been a 14 year process for Gonzaga since coming onto the scene with a surprise Elite Eight run in the tournament) it is not very likely that the Bulldogs climb to the top of the polls will translate to a National Championship. Gonzaga is currently 5th in our Power Rankings, 3rd in offensive efficiency and 20th in defensive efficiency. After being second in our rankings for the last month Indiana (1st offensive efficiency, 18th defensive efficiency) has retaken the top spot over Florida (6th offensive efficiency, 4th defensive efficiency).

Indiana was our mostly likely champion in our first two simulations of Bracketology. Then as Florida began playing better and their seeding improved they became our most likely winner in each of the last two simulations. In the current iteration Florida is a 3-seed and Indiana is still projected as a 1-seed. As always, seeding is just as crucial to a long tournament run as the opponents you will face.

After 50,000 simulations the most likely National Champion is the Indiana Hoosiers. Indiana wins 20.0% of all the simulated tournaments. This number is down from our previous simulation on February 19th when Florida was projected to win 28.8% of the simulations and while Indiana was not our most likely winner they still won more tournaments, 26.8%, than they do currently. The seeding for both Indiana and Florida impacts their odds in this simulation. Indiana drawing Louisville (our 3rd most likely winner) in their region and Florida being a 3-seed greatly diminishes each team's chances. To see how seeding can impact the outcome of the tournament for these teams or your favorite team, try the Bracket Simulator where you can build and simulate any bracket you wish.

With the additional information we have gained from two more weeks of basketball, the championship odds for a number of teams have changed. Below is a look at five teams who have consistently been the favorites to win and how their odds have changed over last the last two months.

  Jan. 8th Jan. 22nd Feb. 5th Feb. 19th Mar. 7th
Indiana 14.1% 16.9% 19.2% 26.8% 20.0%
Duke 10.3% 9.4% 7.9% 2.8% 7.0%
Louisville 13.2% 13.0% 7.3% 8.1% 11.0%
Michigan 8.5% 8.1% 14.2% 6.0% 6.3%
Florida 9.0% 11.9% 21.7% 28.8% 16.9%

Here is a look at each of the above team's resumes for a #1 seed and what opportunities they have left to impress the selection committee before the tournament begins.

Indiana:

Key matchups remaining - @ #7 Michigan, Big 10 Conference Tournament

RPI SOS vs. RPI Top 25 vs. RPI Top 50 Bad Losses
10 15 7-2 7-5 None

Duke:

Key matchups remaining – @ UNC, ACC Conference Tournament

RPI SOS vs. RPI Top 25 vs. RPI Top 50 Bad Losses
1 1 6-1 8-2 Maryland Virginia

Louisville:

Key matchups remaining – vs. #24 Notre Dame, Big East Conference Tournament

RPI SOS vs. RPI Top 25 vs. RPI Top 50 Bad Losses
3 8 3-3 7-4 Villanova

Michigan:

Key matchups remaining – vs. #1 Indiana, Big 10 Conference Tournament

RPI SOS vs. RPI Top 25 vs. RPI Top 50 Bad Losses
8 61 4-3 8-4 Penn State

Florida:

Key matchups remaining - @ Kentucky, SEC Conference Tournament

RPI SOS vs. RPI Top 25 vs. RPI Top 50 Bad Losses
6 29 2-2 4-3 Arkansas Tennessee

In 2011, all 16 teams seeded 1-4 in the bracket were at least 50% likely to make the Sweet 16. As the bracket is currently constructed Oklahoma State (29.1%), Miami FL (48.6%), New Mexico (33.1%), Marquette (42.3%), and Kansas State (29.6%) are all less than 50% likely to win two games in the tournament. By the second weekend we could lose a #2, two #3 and two #4 seeds. Georgetown only has a 52.4% chance of making it to the Sweet 16, which is by far the lowest chance of any #1 seed in any simulation we have ran.

There are five predicted upsets in the first round; there were three in the last simulation. In addition there are a total of ten games in the first round in which the favored team wins 60% of the time or less which lends itself to another wild March Madness.

Conference Breakdown

Big East – 8 teams, Big Ten – 7 teams, Big 12 – 6 teams, Pac-12 - 6, ACC - 5 teams, Mountain West - 5 teams, Atlantic 10 - 5 teams, SEC - 3 teams, West Coast - 2 teams, and Missouri Valley - 2 teams.

Below are teams considered in Bracketology to be On the Bubble. Next to each team is our overall ranking plus where they rate offensively and defensively.

Last Four In Overall Offense Defense RPI SOS
Iowa State 32 8 132 51 53
Temple 68 38 131 41 72
Villanova 51 102 35 53 17
Tennessee 53 48 85 56 48
           
First Four Out Overall Offense Defense RPI SOS
Alabama 59 153 26 60 85
Kentucky 36 29 70 53 78
Southern Miss 62 65 76 42 87
Ole Miss 48 40 82 57 162
           
Next Four Out Overall Offense Defense RPI SOS
Maryland 57 95 48 84 121
Baylor 43 42 63 67 28
Arizona State 75 96 90 93 134
Arkansas 61 88 75 76 69

South Region

Most Likely Final Four team: Florida (40.9%)

Final Four Sleeper: Wisconsin (14.6%)

Cinderella (double-digit seeds going deep): Saint Mary's (32.2% to make Sweet 16)

Chance NCAA Champion is from Region: 26.4%

Three Most Important Players:  Kenny Boynton (Florida), Otto Porter Jr. (Georgetown), Shane Larkin (Miami)

First Round Upset: #10 Saint Mary's over #7 UCLA (60.9%)

Closest First Round Game: #8 NC State over #9 California (59.6%)

West Region

Most Likely Final Four team: Gonzaga (31.7%)

Final Four Sleeper: Ohio State (13.5%)

Cinderella (double-digit seeds going deep): La Salle (13.3% to make to Sweet 16)

Chance NCAA Champion is from Region: 20.6%

Three Most Important Players: Trey Burke (Michigan), Kelly Olynyk (Gonzaga), and Deshaun Thomas (Ohio State)

First Round Upset: None

Closest First Round Game: #7 UNC over #10 Oklahoma (56.0%)

East Region

Most Likely Final Four team: Duke (26.4%) 

Final Four Sleeper: St. Louis (4.5%)

Cinderella (double-digit seeds going deep): Akron (15.7% to make Sweet 16)

Chance NCAA Champion is from Region: 16.8%

Three Most Important Players: Ryan Kelly (Duke), Ben McLemore (Kansas), and Keith Appling (Michigan State)

First Round Upset: #11 Middle Tenn. over #6 Notre Dame (50.6%)

Closest First Round Game: In addition to above, #8 Memphis over #9 Illinois (51.3%)

Midwest Region

Most Likely Final Four team: Indiana (40.1%)

Final Four Sleeper: Creighton (3.7%)

Cinderella (double-digit seeds going deep): Iowa State (13.0% likely to make Sweet 16)

Chance NCAA Champion is from Region: 36.2%

Three Most Important Players: Cody Zeller (Indiana), Peyton Siva (Louisville), and Doug McDermott (Creighton)

First Round Upset: #10 Minnesota over #7 Butler (70.3%)

Closest First Round Game: #9 Creighton over #8 Missouri (52.2%)

Archive

Ad

The Predictalator

The Predictalator is the most advanced sports forecasting software available today because it has the ability to account for all of the relevant statistical interactions of the players (playing or not playing/injured), coaches, officials, fans (home field advantage) and weather in each game.

7/28/2014 Highlight: Using the ResultsFinder for the week of July 21st - 27th, one could find that all featured, half-bet or better MLB plays (including half-bets and normal picks) went 58-37 (59% ML, O/U and RL).

With training camps opening for the NFL, it's also worth noting that the NFL Preview will launch on August 6th, with the college football preview launching a week later on August 13th. For the 2013 season, playable picks for all NFL games went 120-89 (57% ATS), culminating in a playoff run in which all plays went 9-1-1 (90% ATS). Seeing similar success, especially with strongest opinions, the college football Locks of the Week went 14-2 (88% ATS).
Ad
Ad
Ad

Recommended Features