Tag Archives: NCAA bracket

2017 Bracket Predictions

  • The champ will be either Duke, Kentucky, North Carolina, UCLA, or Oregon
  • IF Kansas gets past Miami, they will lose in the Sweet 16 to Iowa State
  • Gonzaga won’t reach the Final Four
  • Extremely over-seeded Minnesota will lose to MTSU in the First Round
  • Extremely over-seeded Maryland will lose to Xavier in the First Round
  • St. Mary’s will pull off the upset of the tournament by beating Arizona in the Second Round

2017 Most Over-seeded Teams

Undeserving 1-seeds have consistently under-performed, and this year North Carolina and Kansas fit the bill.

I am using team ratings from my computer model to determine if a team is over-seeded in the tournament.  1-seeds should be one of the top four teams, 2-seeds should be teams 5-8, and so on.  I’d like to mention that my model’s not too shabby, either.  It performed better than ESPN’s model at picking against the spread.

Historically (going back to 2013), only the top seeds (1 through 6) have demonstrated a clear trend of under-performing when my model says they were over-seeded

Overrated 2-seeds

  • My model says Duke and Arizona both should’ve been 4-seeds
  • 26 of the last 60 2-seeds were over-rated by that much or more.
    • None won the championship
    • Two (including Duke in 2012) lost in the first round
    • Ten lost in the Second Round
    • Only 5 reached the Final Four
  • Duke and Arizona do, however, have the DNA of a championship team

Overrated 3-seeds

  • 3-seeds that should be be 6-seeds or worse have consistently under-performed
  • This year, only Baylor is that overrated
  • There is one outlier – in 2011, Connecticut should’ve been a 6-seed, but they won it all
  • There have been 16 other 3-seeds that overrated
    • None reached the Final Four
    • Two lost in the First Round
    • Seven lost in the Second Round

Overrated 4-seeds

  • Purdue should’ve been a 5 and Butler a 6
  • 27 of the last 60 4-seeds were over-seeded
    • None reached the Championship game
    • Only 2 reached the Elite 8 and one reached the Final Four
    • 8 lost in the First Round
    • 9 lost in the Second Round

Overrated 5-seeds

  • Minnesota should’ve been an 8-seed
  • 24 of the last 60 5-seed have been  as grossly over-seeded as Minnesota or more
    • 13 (over half!) of them lost in the First Round
    • In 2010, Butler surprised everyone by reaching the Championship game as a 5-seed
    • Besides Butler, none of the other over-seeded 5’s in this group reached even the Elite 8

Overrated 6-seeds

  • Maryland shouldn’t have even made it into the tournament
  • 22 of the last 60 6-seeds were significantly over-seeded
    • 13 (59%) of them lost in the First Round
    • None of the reached the Final Four
    • 5 of them actually won multiple games, so all is not lost for Maryland.  Forget about a Final Four run, though.

Over-seeded 1-Seeds Underperform

By undeserving, I’m talking about 1-seeds that my computer model says are not one of the top four teams in the tournament.

This year, my model says North Carolina should’ve been a 2-seed and Kansas should’ve been a 5-seed.

Going back to 2003, my model says 23 out of the 60 1-seeds were over-seeded.

The average number of wins for deserving 1-seeds was 3.8 (4 wins get you to the Final 4), while the average number of wins for undeserving 1-seeds was 2.3.  That’s a big difference!

5 of the 23 undeserving 1-seeds lost in the Second Round, and none of them reached the championship game.

That’s a decent number of data points to indicate my model’s on to something here.

2017 Advice

Definitely don’t pick Kansas to win it all.  Fight the urge!  Remember, the Quarterfinal Rule already declared them ineligible.

The data shows an overseeded 1-seed hasn’t won the championship since 2003, so think twice about picking UNC.

I’m predicting Iowa State to knock Kansas out in the Second Round and UNC to lose in the Elite 8 to… I’m not sure, take your pick between Kentucky, UCLA, and Wichita State.

2017 Bracket Tip Sheet

Pick the Champ First

  • 18 of last 19 champs were a 1-, 2-, or 3-seed
  • This year’s champ will probably be either Arizona, Kentucky, North Carolina, Oregon, UCLA, or Duke.  To see why, check out my post about teams with championship DNA

Next Pick Your Final Four

  • All four 1-seeds have made the Final 4 only once
  • All four 1-seeds failed to reach the Final 4 only twice
  • On average, the Final 4 consists of at least one 1-seed and only one team seeded 4 or worse

Elite Eight

  • On average, the Elite 8 consists of three 1-seeds, two 2-seeds, one 3-seed, and one team seeded lower than 5

The Cinderellas

  • On average, 6 double digit seeds advance to the Second Round.  The average over the last 10 years is 7 teams
  • On average, 2 double digit seeds advance to the Sweet 16
  • Never pick a 16-seed to beat a 1-seed
  • Play it safe, and don’t pick a 15-seed to beat a 2-seed.  It’s too rare to be worth the risk.

The DNA of a National Championship Team

A thorough analysis of past national championship teams has shown that they have certain traits and characteristics in common; trends that like the “Democratic Peace Theory” may not be classified as academic law, but as of yet, the theory hasn’t been broken.  Put the common characteristics together and we have the DNA of a championship team.  Scroll all the way to the bottom of this post if you just want to see a list of this year’s contenders.

Caution:  All championship teams fit this description and have these championship characteristics, but not all teams that fit this description win the championship.  Every year there are between 3 and 10 contenders with the right DNA that don’t win it all and might even suffer an early exit from the tournament (for example, 2016 Michigan State losing in the First Round or Kentucky’s second round exit in 2004).

Out of hundreds of statistical categories and characteristics, the data has shown so far that there are five traits to look for in a potential national champion:  conference affiliation, AP rank trend, conference tournament performance, certain stat thresholds, and tournament seeding.

Conference Affiliation

Since 1991, all but one champion came from one of the following conferences:

  • ACC
  • Big 12
  • Big East
  • Big Ten
  • PAC 12
  • SEC

The lone exception is when Connecticut won the title in 2014.  That was their first season in the AAC after leaving the Big East.  The data seems to indicate that there is strong correlation between success in the NCAA Tournament format and a season long schedule that is formidable day in and day out.  Whether that is a hardening of the competitor’s mental fortitude, financial resources to better condition and prepare athletes, or a myriad of other possibilities is up for debate in the comments section below.  The data does not demonstrably elucidate the reason behind this correlation, just that there is one.

AP Rank Trend

All but two of the past 28 champs were ranked no worse than 12th in the AP Poll seven weeks before the tournament began.  Just like in football, attempting to accurately rank a team at the beginning of the season is a futile exercise–that early fall darling can turn out to be as ugly as your first prom date.  However, give them a bit of time to play and produce some data to be analyzed and we can form a much better conclusion as to the quality of that team and their opponents.  In 2017, those teams are:

  • Villanova
  • Kansas
  • Gonzaga
  • Kentucky
  • Baylor
  • Florida State
  • Arizona
  • UCLA
  • North Carolina
  • Oregon
  • Butler
  • Virginia

For teams like Kentucky, Baylor, Florida State, Butler, and Virginia, who have dropped a bit in the poll since then, read on.  They aren’t dead but they have some extra work to do.

Conference Tournament Performance

There are three significant points to consider from conference tournament results:  the Quarterfinal Rule, how contenders slipping in the polls perform, and contenders that fail to reach their conference tournament final.

The Quarterfinal Rule

There is a pretty fascinating trend I call the Quarterfinal Rule that helps eliminate teams from the candidate pool.  The Quarterfinal Rule states that if a team fails to advance past the quarterfinals in their conference tournament, they will not win the national title.  So far, so-called contenders that have found themselves on the wrong side of the Quarterfinal Rule are:

  • Kansas
  • Baylor
  • Louisville
  • Virginia
  • Butler
  • Purdue
  • Florida

Performance of contenders slipping in the polls

There have been 3 national champs that were ranked in the top 12 (as explained above), but dropped a bit in the polls leading up to the tournament.  They each won their conference tournament (UConn in ’04 and ’11, and Florida in ’06).  In 2017, the teams that need to win their conference tournaments are:

  • Kentucky
  • Baylor (already lost)
  • Florida State (already lost)
  • Butler (already lost)
  • Virginia (already lost)

Contenders not reaching the conference tournament final

Even if a team loses in their conference semifinal, they might still be okay.  Since 1986, only 6 national champions failed to reach their conference tournament finals.  Five of the 6 were ranked in the top 4 of the AP poll, and still received 1-seeds (Arkansas in ’94, Maryland in ’02, North Carolina in ’05 and ’09, and Duke in ’15).  The other team, Syracuse in ’03, rose in the AP poll from 24th to 13th in the weeks leading up to the NCAA Tournament.  In 2017, teams that are not eliminated from contention even though they lost in their conference semifinals are:

  • UCLA
  • North Carolina

Stats

Okay, this one uses some customized analytics I developed.  Long story short, I took key team stats, adjusted them to account for opponent strength, and normalized them to create a score for each statistical category.  So far, I only have all the regular season data I need back to the 2002-2003 season.  However, this has still allowed for some interesting and fairly solid conclusions.  Since 2003, every champion scored well in 4 statistical categories:  A) Overall team rating (according to my model), B) Field goal percentage, C) Total rebounds, and D) Points per game.  Here are this year’s teams that meet the champion thresholds for these stats (in alphabetical order):

  •  Arizona
  • Cincinnati
  • Duke
  • Florida
  • Florida State
  • Gonzaga
  • Kansas
  • Kentucky
  • Louisville
  • North Carolina
  • Oregon
  • Purdue
  • SMU
  • UCLA
  • West Virginia
  • Wichita State

Tournament Seeding

The good-natured folks on the selection committee invariably make seeding mistakes and overlook valid resumés.  However, don’t dismiss their ability to determine the very best teams.  28 of the past 32 champions have been a 1-, 2-, or 3-seed.  Since 1998, only one champ wasn’t a 1-, 2-, or 3-seed (7-seeded Connecticut in 2014).  This year’s 1-, 2-, and 3-seeds:

  1. Villanova, Kansas, North Carolina, Gonzaga
  2. Duke, Arizona, Kentucky, Louisville
  3. Baylor, UCLA, Oregon, Florida State

So…This Year’s Contenders Are…

I’ll remind everyone that these conclusions aren’t meant to be debated like the old hat baseball experts in Moneyball, “Yah, he he’s got a baseball body, he looks real good!” That is emotion speaking.  This is the result of an analysis of thousands of data points to identify trends and historical precedence that give confidence in the probability of an event happening in the future.  Something crazy could happen, but the numbers say it most likely won’t.

Taking all of these elements into consideration there are only six teams that have championship DNA this year:

  • Arizona
  • Kentucky
  • North Carolina
  • Oregon
  • UCLA
  • Duke (see below)

But…

…’cause there’s always a but…  I included Duke even though they weren’t ranked in the top 12 in the right week, which is the weakest requirement of the five.  Duke compares favorably with the only two champs that also didn’t meet the AP rank threshold -2003 Syracuse and 2014 Connecticut. Given Duke played a significant stretch of their season without their Hall-of-Fame head coach AND they killed it in the ACC tournament, I’ve included Duke as one of the teams that has what it takes to win it all.

The only thing keeping Gonzaga off the list is their conference affiliation.  It sure is tempting to include Gonzaga in the list of contenders, as their impressive resume has been discussed ad nauseam.  Statistically, they are an elite team.  They have maintained an elite position in the polls, won their conference tournament, and were awarded a 1-seed.  Perhaps they have what it takes, but their win would be historic.  The 2006-2009 Memphis teams were remarkably similar – very talented, well ranked, measured up statistically, won their conference tournaments, and received 1- or 2-seeds.  In 2008, they even reached the Championship game, but still were unable to pull it off.  In 2013, we saw Gonzaga in this exact scenario.  Besides their conference affiliation, they had all the other characteristics of a championship team.  The 1-seeded Zags lost in the second round.

So, those are the numbers.  Now the debate begins in the comments section below.