Bracket Analysis - 2018
2018 Tournament Pages:
Prediction Machine has already played the 2018 NCAA Tournament 50,000 times to create the Bracket Odds table with the chances of any team making it to any level in the tournament, as well as the Predictalated Bracket with the machine's most likely outcomes. Here, we break down each region to look at the best teams, sleepers and Cinderellas.
How it Works
Prediction Machine 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.
Virginia, the top overall seed, is our most likely title winner at 19.4% likely despite playing in the toughest overall region. That percentage is almost identical to our most likely champion at this time last year (Villanova at 19.8%). In 2015, Kentucky was the most likely champion we have ever seen at 44.3% (Duke and Wisconsin were next). In 2014, no team won the title more than 11% of the time. This year may seem "wide open" in a relative sense, but it is right in the middle based on the previous 14 seasons' simulations.
Just three teams this year - Virginia, Villanova (17.9%) and Duke (17.0%) are more than ten percent likely to win it all, while there are only four other teams - UNC (9.7%), Purdue (6.5%), Michigan State (6.4%) and Cincinnati (5.6%) that have a five percent chance or better.
Duke, which entered the season as the number one team in the nation and battled through injuries, is in an interesting spot. The Blue Devils, likely the most talented starting five in the country, rank second overall behind Virginia in our final Power Rankings and would be our most likely champion if they switched spots wtih Virginia (or Villanova or Xavier) in the bracket. In its current region, Duke may have to face Michigan State (#4 in or Power Rankings) and Kansas (#7)
It has already been a wild 2017-18 season. We saw more AP Top Ten teams lose to unranked opponents than ever before and at various points, Arizona State and Oklahoma (two double-digit, bubble teams) were condiered one seeds. In a season that has been as unpredictable as this one, how wide open is the NCAA tournament going to be? To answer that question we looked back at the last eight tournaments on PredictionMachine.com (2010 to 2017) we simulated and averaged the probability for each of the top ten teams to cut down the nets. Then we compared that average to the ten most likely 2018 champions as of right now.
How much of a difference is there heading into March Madness?
|Title Chances Rank||Average 2010-2017||2018 Simulations|
There are ten teams with at least a 2.3% chance to win the championship and that top ten is roughly 90% likely to do so. Despite as wide open as the season felt, the bracket is mostly average aside from being just a little more top heavy than we usually see. It make some sense as early surprise teams in the AP rankings fell off throughout the year and there is not much depth among potential contenders in each conference (most power conferences have 2-3 really good teams and then several that are decent, yet interchangeable)
The most likely Final Four for 2018 is (in order of likelihood) Virginia, Villanova, Duke and UNC. That is not all one seeds, but it is three ACC teams in the Final Four. It's important to keep in mind that this exact Final Four only happens 4.0% of the time (last year's most likely Final Four was 1.7% likely). Purdue is next, while only Michigan State, Cincinnati, and Kansas have a better than one-in-five chance at the Final Four otherwise.
In the most likely bracket, a quick review of the results may seem pretty chalky. The most likely Elite Eight has three one seeds and four two seeds and one four seed (Gonzaga). Furthermore, one five seed (West Virginia) makes the Sweet 16 with all other regions seeing 1-4 seeds make the most likely Sweet 16. And, on the flip side to that, we predict one 12 seed and two 11 seeds to win outright in the First Round. After West Virginia, however, most seeds from 5 through 12 look pretty interchangeable.
While we would actually prefer to reward resume over talent in seeding and region placement ("most deserving" over "best" is what committee should review, while Prediction Machine deciphers "best"), some reasonable consideration needs to be given for talent and expected chances for success. For instance, in our updated Power Rankings, Nine teams got into the tournament as at-large teams despite being ranked 45th or worse. Those teams are St. Bonaventure (#80), Kansas State (#66), Arizona State (#58), Oklahoma (#56), Providence (#52), Syracuse (#50), Arkansas (#48), NC State (#47) and Florida State (#46). Last year, five such teams made the tournament, so this is not necessarily better.
This year, there are seven teams in the top 45 in our final power rankings of the best teams in college basketball that did not make the tournament. Those are Notre Dame (#19), Saint Mary's (#33), Baylor (#34), Penn State (#35), Louisville (#39), USC (#41) and Marquette (#45).
Three regions are very even at the top. The South region has the highest chance to win the championship at 28.5% and also has five teams in the top 13 of our Power Rankings. The Midwest (28%) and East (27%) are very close in championship chances. The West, which has the weakest one seed, is just 16.5% likely to win it all.
After Virginia, Villanova, Duke, UNC, Purdue, Michigan State and Cincinnati, the list of teams with at least a one percent chance to win the championship also includes (in order): Kansas, Gonzaga, Michigan, Arizona, Xavier, Tennessee and West Virginia. The group of teams with a greater than one percent chance at a title includes three teams from the ACC and Big Ten, two teams from the SEC, Big East and Big 12, on team from the Pac-12 and one team from the West Coast conference.
West Virginia, Kentucky, Houston, Florida, Ohio State and Missouri (the best 5+ seeds with respect to chances to make the Final Four) may be sleeper contenders. They do not really qualify as Cinderella candidates (double-digit seeds who could make a deep run into the tournament).
There are some double-digit seeds with decent chances of making the Sweet 16 including: ten seeds Texas and Butler at 14.8% and 14.5% respecitvely to make Sweet 16, 11 seed Loyola (IL) (14.1%), 12 seed Davidson (12.0%) and 11 seeds UCLA and San Diego State (11.1% and 9.6% respectively). The strongest 13+ seeds are Charleston (16.6% likely to win a First Round game), Buffalo (16.0%), Stephen F. Austin (15.4%), UNC-Greensboro (14.0%) and Marshall (13.1%). No 15 or 16 seed has a better than six percent chance to win a game after the Opening Round. The likelihood of all four one seeds winning in the first round yet again is 89.6%.
There are just three predicted upsets in the Round of 64 (outlined below), yet a total of 12 games in which the favored team wins two thirds of the time or less. The most likely 12-over-5 upset is New Mexico State with a 33.8% chance to win over Clemson.
Most Likely Final Four team: Virginia (49.5%)
Final Four Sleeper: Kentucky (5.1%)
Teams in Top 25 of Power Rankings: 5
Cinderella (double-digit seeds going deep): Loyola (14.1% to make Sweet 16 - the South has 3 of 4 most likely double digit seeds to make Sweet 16)
Chance NCAA Champion is from Region: 28.5%
Three Most Important Players: Shai Gilgeous-Alexander (Kentucky), Jacob Evans (Cincinnati) and Admiral Schofield (Tennessee)
First Round Upset: #10 Texas over #7 Nevada (51%)
Closest First Round Game: In addition to above, #6 Miami over #11 Loyola (53.2%)
Most Likely Final Four team: Villanova (45.9%)
Final Four Sleeper: West Virginia (8.0%)
Teams in Top 25 of Power Rankings: 6
Cinderella (double-digit seeds going deep): Butler (14.5% to make to Sweet 16)
Chance NCAA Champion is from Region: 27.0%
Three Most Important Players: Mikal Bridges (Villanova), Collin Sexton (Alabama) and Keenan Evans (Texas Tech)
First Round Upset: #9 Alabama over #8 Virginia Tech (54%), #10 Butler over #7 Arkansas (57.9%)
Closest First Round Game: In addition to above, #6 Florida over #11 UCLA/St. Bonaventure (67.8%)
Most Likely Final Four team: UNC (39.1%)
Final Four Sleeper: Gonzaga (17.6%)
Teams in Top 25 of Power Rankings: 7
Cinderella (double-digit seeds going deep): San Diego State (9.6% to make Sweet 16)
Chance NCAA Champion is from Region: 16.5%
Three Most Important Players: Michael Porter (Missouri), Trevon Bluiett (Xavier) and Keita Bates-Diop (Ohio State)
First Round Upset: None
Closest First Round Game: #7 Texas A&M over #10 Providence (59.3%)
Most Likely Final Four team: Duke (45.2%)
Final Four Sleeper: Seton Hall (1.5%... top three seeds are 90%+ to make Final Four)
Teams in Top 25 of Power Rankings: 6
Cinderella (double-digit seeds going deep): New Mexico State (14.4% likely to make Sweet 16)
Chance NCAA Champion is from Region: 28.0%
Three Most Important Players: Trae Young (Oklahoma), Udoka Azubuike (Kansas) and Cassius Winston (Michigan State)
First Round Upsets: None
Closest First Round Game: #7 Rhode Island over #10 Oklahoma (53.0%)
2018 Tournament Pages:
Tournament History Highlight: Results have been consistently dominant in the NCAA Tournament in years past. Over 8 NCAA Tournaments at PredictionMachine.com, featured, "normal" or better picks (anything with greater than 57% confidence) are 76-40 (66% ATS), while the top play each day is 57-30 (66% ATS). This includes being 39-18 (67% ATS) on normal or better picks and 14-2 on Top Plays of the Day all-time in the First Round (first Thursday and Friday each tournament)
Picks and predictions on every college basketball game throughout the entire tournament, as well as every regular season NBA, NHL and MLB game (with win total picks), including pick confidence and bet value recommendations on every play are now available at a discount. See more.