Madness Explained: A Comprehensive Overview of Why Teams Make the Division I Men’s Basketball Final Four

8583897140_dfe6223e09_z

Bryan Fuller / MGoBlog

WHAT DOES ALL THIS MEAN?

NOTE: THIS STUDY IS ACCOMPANIED BY AN EXCEL SPREADSHEET LISTING ALL FINAL FOUR TEAMS FROM 2002-2016.

First and foremost, the only teams listed in this study and its accompanying spreadsheet are teams that made the Final Four (the champions are in bolded text in the spreadsheet). The figures of “Efficiency Margin” (EM), “Adjusted Offensive Efficiency (AdjOE)”, “Adjusted Defensive Efficiency” (AdjDE), “Luck”, “Adjusted Tempo (AdjTempo)” and “3PA/FGA” are all pulled from the database of Ken Pomeroy. Here are links both to his website and to an explanation of what the aforementioned statistics measure:

Ken is a preeminent mind in the field of basketball analytics and has created tempo-free and opponent quality-adjusted formulas that measure team quality in Division I Men’s Basketball. Unless otherwise stated, the teams in this particular study range from 2002-2016 exclusively because those are the only years in his database. All of the data used (except for Luck and 3PA/FGA, which were unavailable) was recorded pre-NCAA tournament and thus provides a valid analog to the numbers we will use in March to fill out our brackets.

“3PA/FGA” lists a team’s percentage of shots that are three-pointers (makes and misses alike), “Winning Streak” lists a given team’s winning streak heading into the NCAA tournament, “Preseason AP Poll” lists their ranking in that year’s preseason AP poll and “Regular Season Title” and “Conference Tournament Titles” list whether the teams in question won their regular season or conference tournament titles.

When “ranks” are discussed (i.e. West Virginia has a Luck rank of 287), the numbers in question are not the values the stat measures but the national rank. I have all but excluded most of the actual values of the statistics in the written portion of the study because I feel as though the rank makes for much more digestible context (if a person on the street told me that UCLA had an AdjOE of 125.3, I’d smile blankly; telling me they have the most efficient offense would resonate much more).

Throughout this examination, various teams are deemed to be “Final Four Contenders” or “Championship Contenders”; these classifications are based on an interpretation of both objective factors (their statistical profile, their projected seed line) and subjective factors (the public consensus, how much talent a team has). Some teams may be excluded that you feel deserve recognition.

Additionally, at the end of each section I offer takeaways in the form of “Ensure that your Final Four selections … “. The level of certainty or confidence that you should attach to each section is solely under your purview, but I recommend referencing the statistics within each section when determining certainty. For example, if 55 out of 60 Final Four teams share a particular characteristic, I would give credence to the importance of that section over one featuring a characteristic that only 44 out of 60 share. That being said, the final decision ultimately lies with you.

Without further adieu, here are some conclusions that can be drawn from the attached data:

SEEDS

I will defer this particular explanation to Shane Ryan, formerly of Grantland. In his 2015 article “MoneyBrackets: A 15-Step Guide to Conquering March Madness”, he wrote: “… we can see that since the tournament expanded to 64 teams, 74 of the 120 Final Four teams have been either 1- or 2-seeds. That’s 61.7 percent, and while it may seem like those numbers argue for picking two or three 1- or 2-seeds, consider this: In those same years, there have been 14 3-seeds and 13 4-seeds in the Final Four. Combine them, and that’s not even one 3- or 4-seed in the Final Four per year. Even if one does make it in 2015, your odds of actually picking the right one are, at best, 1-in-8. Go any lower in the seeds, and your odds get exponentially worse.”

Through 2016, 80 of the 128 Final Four teams (62.5%) have been 1- or 2-seeds, so the number of top-heavy semifinalists has actually increased in the two years since Shane penned his article. In sum, there’s roughly a 37% chance a seed lower than 1 or 2 will make the Final Four in a given year, but there’s a 12.5% chance or less you’ll correctly identify which team will do so.

ADVICE: I recommend you choose only 1- or 2-seeds for your Final Four unless there is an exceptionally qualified 3-, 4- or 5-seed paired with a particularly weak 1- and 2-seed (these historical anomalies will be discussed later).

 

EFFICIENCY MARGIN

With regard to Efficiency Margin (EM for short), here are the historical trends:

  • 46 out of 60 Final Four teams (76.6%) had an EM equal to or greater than a rank of 15 in a given year.
  • 41 out of 60 Final Four teams (68.3%) had an EM equal to or greater than a rank of 10 in a given year.
  • 27 out of 60 Final Four teams (45%) had an EM equal to or greater than a rank of 5 in a given year.
  • 12 out of 15 champions (80%) had an EM equal to or greater than a rank of 6 in a given year. The three that did not – 2003 Syracuse (20), 2011 UConn (16) and 2014 UConn (25) – were historical outliers with regard to winning the championship in more ways than one (see “Outliers” for more information).

ADVICE: While I advise ensuring that each of your 4 Final Four teams has an EM rank of at least 15, the most significant portion of this segment involves ensuring that your eventual champion adheres to a rank of 6 or higher in Efficiency Margin.

NOTABLE TEAMS CURRENTLY IN DANGER (3/12/17):

FINAL FOUR CONTENDERS:

Oregon – EM rank: 16

UCLA – EM rank: 18

Arizona – EM rank: 20

CHAMPIONSHIP CONTENDERS:

Kansas – EM rank: 10

Duke – EM rank: 12

 

ADJUSTED OFFENSIVE EFFICIENCY

With regard to Adjusted Offensive Efficiency (AdjOE for short), here are the historical trends:

  • 13 out of the past 15 Final Fours (86.6%) have featured at least 1 team who entered the tournament ranked in the top 5 of AdjOE (6 of the 15 Final Fours had 2 teams).
  • 54 out of 60 Final Four teams (90%) had an AdjOE equal to or greater than a rank of 50 in a given year.
  • 52 out of 60 Final Four teams (86.6%) had an AdjOE equal to or greater than a rank of 40 in a given year.
  • 44 out of 60 Final Four teams (73.3%) had an AdjOE equal to or greater than a rank of 20 in a given year.

ADVICE: Ensure that your Final Four selections have an AdjOE rank of at least 50.

NOTABLE TEAMS CURRENTLY IN DANGER (3/12/17): NONE

 

ADJUSTED DEFENSIVE EFFICIENCY

With regard to Adjusted Defensive Efficiency (AdjDE for short), here are the historical trends:

  • 54 out of 60 Final Four teams (90%) had an AdjDE equal to or greater than a rank of 40 in a given year.
  • 49 out of 60 Final Four teams (81.6%) had an AdjDE equal to or greater than a rank of 30 in a given year.
  • 39 out of 60 Final Four teams (65%) had an AdjDE equal to or greater than a rank of 20 in a given year.

ADVICE: Ensure that your Final Four selections have an AdjDE rank of at least 40.

NOTABLE TEAMS CURRENTLY IN DANGER (3/12/17):

UCLA – AdjDE rank: 78

 

LUCK

In tackling what “Luck” measures, we begin with an explanation from its creator, the aforementioned Pomeroy: “The easiest one to understand is Luck, which is the deviation in winning percentage between a team’s actual record and their expected record using the correlated gaussian method. The luck factor has nothing to do with the rating calculation, but a team that is very lucky (positive numbers) will tend to be rated lower by my system than their record would suggest.”

Essentially, “Luck” measures how a team’s actual record stacks up to their expected record based on their performance. An example of a “lucky” team in the 2016-2017 season is Kansas, who has an EM rank of 8 but a 25-3 record – better than all but two teams in front of them in the EM rankings (correspondingly, their Luck rank is 15). The quintessential “unlucky” team in 2016-2017 is West Virginia, who has an EM rank of 3 paired with a 22-6 record (their Luck rank is 287).

The crux of the ambiguity with the Luck rating is that it’s not exactly clear what makes a particular team lucky or unlucky. It could be any combination of shooting, coaching, fouls, or, as its name implies, blind serendipity (or a decided lack thereof).

A second sticking point with this measure exists in the fact that we only have access to post-tournament Luck ratings; conversely, every other Pomeroy rating this study uses was taken before the NCAA tournament was played in a given year. However, I am confident that, especially late in a season, Luck does not change so dramatically as to become an unusable or invalid figure over a span of only six games.

Vagueness and uncertainty aside, some historical trends have still emerged with regard to Luck. They are as follows:

  • No champion has finished with a Luck rank in the 300s, and only 1 (2013 Louisville, who finished at 229) has finished with a Luck rank in the 200s.
  • 50 out of 60 Final Four teams (83.3%) had Luck equal to or greater than a rank of 199 in a given year.
  • 22 out of 60 Final Four teams (36.6%) had Luck equal to or greater than a rank of 99 in a given year.
  • 8 out of 60 Final Four teams (13.3%) had Luck equal to or greater than a rank of 50 in a given year; only 5 teams have ever finished with a rank of 30 or greater, and only 1 has ever finished in the top 10 (2009 Michigan State).

A quick follow-up to the problems with using post-tournament Luck – while it logically follows that teams that go to the Final Four will have higher Luck rankings than teams that don’t (because they win and climb higher in the rankings, while others lose and fall), it is telling that only 1 champion out of 15 ever finished 200 or lower; even if a team started the tournament in the low 200s and won by a close margin all six games, the relatively static nature of the luck rankings given a full season’s worth of data would make it unlikely for a team to move up 100 spots.

Although this is, admittedly, a more conjectural conclusion than I would like, it seems as though there is a “sweet spot” of sorts with Luck; a team does not want to start the tournament too low (~250th or lower) nor too high (~40th or greater).

ADVICE: As such, I would ensure that your Final Four selections do not start with a luck ranking of ~280 or lower OR ~40 or greater.

NOTABLE TEAMS CURRENTLY IN DANGER (3/12/17):

FINAL FOUR CONTENDERS:

Virginia – Luck rank: 298

Florida – Luck rank: 292

Arizona – Luck rank: 37

CHAMPIONSHIP CONTENDERS:

West Virginia – Luck rank: 290

Kansas – Luck rank: 25

 

ADJUSTED TEMPO

With regard to Adjusted Tempo (how many possessions a team plays in a given game), here are the historical trends:

  • 56 out of 60 Final Four teams (93.3%) had an AdjTempo equal to or greater than a rank of 299 in a given year.
  • 22 out of 60 Final Four teams (36.6%) had an AdjTempo rank between 299 and 200, 16 out of 60 (26.6%) between 199 and 100, and 18 out of 60 (30%) between 99 and 1. Of the most recent 28 Final Four teams, only 4 (14.2%) have had an AdjTempo equal to or greater than 99 (3 of these 4 have occurred in the past two years – 2016 North Carolina, 2016 Oklahoma and 2015 Duke, though all 3 finished with Tempo ranks in the 90s or lower, indicating a slow-down of sorts during the tournament itself).

ADVICE: Ensure that your Final Four selections have an AdjTempo rank equal to or greater than a rank of 299.

NOTABLE TEAMS CURRENTLY IN DANGER (3/12/17):

FINAL FOUR CONTENDERS:

Virginia – AdjTempo rank: 351

SMU – AdjTempo rank: 330

Baylor – AdjTempo rank: 337

CHAMPIONSHIP CONTENDERS:

Villanova – AdjTempo rank: 323

 

3PA/FGA

The importance of this particular stat was brought to my attention by Ed Feng, founder of the website The Power Rank and author of the book, “How to Win Your NCAA Tournament Pool”. Ed has his own college basketball rankings and I highly recommend you use them alongside Pomeroy’s in assessing the quality of college basketball teams. Full disclosure: I purchased his book in March of 2015 and recommend you do so as well, as it is packed with insightful information and strategies regarding winning tournament pools.

Teams that rank high in 3PA/FGA take a larger percentage of their shots from three-point distance than most – Michigan, under John Beilein, is a traditional example. Teams that rank low in this stat take two-point shots at a much higher clip than three-point shots; Roy Williams’ North Carolina teams normally embody this particular philosophy, and are therefore usually in the 300s in this stat.

With regard to 3PA/FGA, here are the historical trends:

  • 50 out of 60 Final Four teams (83.3%) had a 3PA/FGA rank equal to or lower than 100.
  • 15 out of 60 Final Four teams (25%) had an 3PA/FGA rank in the 300s, 13 out of 60 (21.6%) were between 299 and 200, and 22 out of 60 (36.6%) were between 199 and 100. Of the 10 teams that finished with a rank in the top 99 of 3PA/FGA and made the Final Four, 5 were 4-seeds or lower, indicating that their presence in the Final Four was a statistical anomaly in the first place (as discussed in the “Seeds” section).

 There has only been one champion who finished in the top 99 of 3PA/FGA: 2016 Villanova. The Wildcats came into the tournament in the top 30 in 3PA/FGA and shot them at a middling clip (34.3%, good for 185th in the country); the combination of these figures, their difficult region and the 2nd round exit they sustained the year prior made them an unpopular pick to make the Final Four. However, in an unforeseeable turn of events, Villanova completely altered their offensive profile throughout their 6 game run through the tournament, shooting 3s at a rate that would have ranked them at 159 in 3PA/FGA and making them at a 50% clip (a nearly 16 percentage point jump from their regular season average). This study opines that this transformation was neither predictable (in 2015, with 6 of the same players that would win the title a year later, the Wildcats had a 3PA/FGA figure of 42.9 for the season (rank: 22) and 42.3 for the tournament alone) nor repeatable (2017 Villanova is currently ranked 25th in 3PA/FGA at 43.9 and features only one rotation player who has not shot at least 50 three-pointers this season). Thus, it is treated as a statistical curiosity and nothing more (for now).

ADVICE: In sum, ensure that your Final Four selections have a 3PA/FGA rank equal to or lower than 100, and do not pick a champion that ranks 100 or higher in 3PA/FGA.

NOTABLE TEAMS CURRENTLY IN DANGER (3/12/17):

CHAMPIONSHIP CONTENDERS:

Villanova – 3PA/FGA rank: 25

 

WINNING STREAK

This stat may be one of the most straightforward with regard to its explanation: simply put, Winning Streak (WS) measures a team’s winning streak heading into the NCAA tournament and nothing else. As such, every team that does not win its conference tournament automatically enters the tournament with a WS of 0.

With regard to WS, here are the historical trends:

  • 55 out of 60 Final Four teams (91.6%) and 15 out of 15 champions had a WS of 10 or less.
  • Since 2002, 25 1- or 2-seeds have had a WS of 8 or higher. 7 of 25 have made the Final Four (28%), and only one, 2013 Louisville, has won the championship.
  • 43 out of 60 Final Four teams (71.6%) and 13 out of 15 champions (86.6%) had a WS or 5 or less.

ADVICE: Ensure that your Final Four selections and your champions have a WS of 10 or less. Within that constraint, a WS of 5 or less is preferable, though not essential.

NOTABLE TEAMS CURRENTLY IN DANGER (3/12/17):

FINAL FOUR CONTENDERS:

SMU – WS: 16

CHAMPIONSHIP CONTENDERS:

Kentucky – WS: 11

 

PRESEASON AP POLL

With many (most) things in life, the more information we are given, the better chance we have of rationally assessing our options and making coherent decisions. However, in the world of college basketball, it is the “semi-informed” state provided by the annual preseason AP poll (PAP) that serves as a superior predictive tool to NCAA tournament success than the weekly, “fully-informed” in-season updates that dot the landscape.

Why is this the case? Take it away, Ken Pomeroy: “It’s when the games start being played that things fall apart. Then, each voter’s bias becomes the same. You lose, you drop in the poll. You win, and you move up.  (I exaggerate slightly. Later in the season at the bottom of the poll, there’s some flexibility. I’m primarily referring to the top ten or so teams here.)

However, in the preseason, voters are free from such restrictions. With every team at 0-0, there is also no conflict in voting a team with a worse record over one with a better record, another thing that mid-season voters try to avoid. With the voters having to use their hoops expertise as opposed to adhering to certain conventions, you end up with an accurate picture of which teams are truly the best.”

This seemingly illogical phenomenon is borne out by the fact that, since 1990, the preseason #1 has won the national title 6 times (compared to only 3 for the team ranked #1 at the end of the regular season), and made the national title game 10 times (compared to only 6 for the latter).[1]

With regard to the preseason AP poll, here are the (other) historical trends:

  • 50 out of 60 Final Four teams (83.3%) were ranked in the preseason AP poll.
  • 36 out of 60 Final Four teams (60%) were ranked in the top 10 of the preseason AP poll.
  • 26 out of 60 Final Four teams (43.3%) were ranked in the top 5 of the preseason AP poll.
  • 9 out of 15 champions (60%) were ranked in the preseason top 5, and 12 out of 15 champions (80%) were ranked in the preseason top 20.

ADVICE: Ensure that your Final Four selections were ranked in the preseason AP poll, and ensure that your champion was at least ranked in the top 20.

NOTABLE TEAMS IN DANGER (3/12/17):

FINAL FOUR CONTENDERS:

Florida – PAP: NR

SMU – PAP: NR

Baylor – PAP: NR

 

REGULAR SEASON AND CONFERENCE TOURNAMENT TITLES

Another section, another deferral – we go again to Shane Ryan, this time at Tobacco Road Blues, who, in a 2015 article titled “Regular Season and Conference Tourney Champions, Since 2000”, concluded: “In terms of making the championship game, winning neither the conference regular season or tournament is just as valuable as winning only the tournament. In terms of making the Final Four, twice as many teams who win neither title make it as teams who won only the conference tournament. When picking your brackets, avoid trendy conference tournament titleists.”

The article in question examines the period of time from 2000-2015. Allow us to examine the results for Final Four teams from 2002-2016:

  • 44 out of 60 Final Four teams (73.3%) won at least one title (regular season or conference tournament).
  • 18 out of 60 Final Four teams (30%) won only their regular season title; 7 out of 60 Final Four teams (11.6%) won only their conference tournament.
  • 19 out of 60 Final Four teams (31.6%) won both titles; 16 out of 60 Final Four teams (26.6%) won neither.
  • 13 out of 15 champions (86.6%) won at least one title; 4 out of 15 (26.6%) won both. The two exceptions were 2014 UConn and 2015 Duke (see “Outliers” for more information).

ADVICE: Give extra credence to teams that win at least one title (value regular season titles over tournament titles by a magnitude of 3 to 1), and ensure that your champion wins at least one title.

NOTABLE TEAMS IN DANGER (3/12/17):

FINAL FOUR CONTENDERS:

Virginia

Florida

Baylor

UCLA

CHAMPIONSHIP CONTENDERS:

West Virginia

Louisville

 

FINAL FOUR CRITERIA

All of the information discussed has lent itself to the creation of rubrics that can be used to decipher both Final Four and Championship contenders. As this is not an exact mathematical formula but a combination of isolated factors, I recommend ensuring that all of your Final Four selections meet either all or all but one of the following criteria. Subjective accounting is encouraged when evaluating teams on this criteria; I’ve penalized Kansas for having an extremely high Luck rank (refer to the “Luck” section above for more details) and Duke for being only two spots above the AdjDE threshold even though, technically speaking, both teams satisfy the criteria.

CRITERIA:

– AdjOE rank of at least 50 or greater (54 out of 60 have met)

– AdjDE rank of at least 40 or greater (54 out of 60 have met)

– Luck rank of at least 199 or greater (50 out of 60 have met)

– AdjTempo rank of at least 299 or greater (56 out of 60 have met)

– 3PA/FGA rank of at least 100 or lower (50 out of 60 have met)

– WS of 10 or less (55 out of 60 have met)

– Ranked in preseason AP poll (50 out of 60 have met)

Before we analyze the current crop of contenders, the historical trends:

  • Overall, 50 out of 60 Final Four teams (83.3%) met all or all but one of the criteria.
  • 24 out of those 50 (48%) met all of the criteria.
  • Among 1 and 2 seeds (120 total) from 2002-2016, 24 teams (20%) met all criteria. 12 of those 24 (50%) went to the Final Four.
  • Also among 1 and 2 seeds from 2002-2016, 30 teams (25%) did not meet 2 or more criteria – only 3 of those 30 (10% – 2014 Wisconsin, 2014 Florida, 2015 Wisconsin) went to the Final Four.
  • Among all 1 and 2 seeds from 2002-2016, 13 teams (10.8%) did not meet 2.5 or more criteria – none of them went to the Final Four.
  • 2 out of 60 teams (3%) have made the Final Four after not meeting 2.5 or more criteria – 2016 Syracuse and 2011 VCU.
Listed below are all teams seeded 1-11 (a 12 seed or lower has never made the Final Four) and the amount of criteria they do NOT meet. As such, 0 is excellent, 1 is passable, and 2 or more is problematic.
FINAL FOUR CONTENDERS (3/12/17):
Gonzaga – 0
Villanova – 2 (3PA/FGA rank of 25, AdjTempo rank of 323)
North Carolina – 0
Kentucky – 1 (WS of 11)
West Virginia – 1 (Luck rank of 290)
Louisville – 0
Virginia – 2 (Luck rank of 298, AdjTempo rank of 351)
Wichita State – 2 (WS of 15, unranked in preseason AP poll)
Florida – 2 (Luck rank of 292, unranked in preseason AP poll)
Kansas – 0.5 (Caution required with Luck rank of 25)
SMU – 3 (AdjTempo rank of 330, unranked in preseason AP poll, WS of 16)
Duke – 0.5 (AdjDE rank is close at 39)
Baylor – 2 (AdjTempo rank of 337, unranked in preseason AP poll)
St. Mary’s – 2 (AdjTempo rank of 350, 3PA/FGA rank of 54)
Purdue – 0
Oregon – 0.5 (3PA/FGA rank is close at 106)
Iowa State – 2 (AdjDE rank of 43, 3PA/FGA rank of 98)
UCLA – 1 (AdjDE rank of 78)
Florida St. – 1 (Unranked in preseason AP poll)
Arizona – 0.5 (Caution required with Luck rank of 37)
Michigan – 4 (AdjDE rank of 69, 3PA/FGA rank of 17, AdjTempo rank of 339, unranked in preseason AP poll)
Cincinnati – 2 (AdjTempo rank of 327, unranked in preseason AP poll)
Wisconsin – 1 (AdjTempo rank of 333)
Oklahoma State – 3 (AdjDE rank of 133, Luck rank of 309, unranked in preseason AP poll)
Notre Dame – 3 (AdjDE rank of 58, 3PA/FGA rank of 69, unranked in preseason AP poll)
Butler – 2.5 (AdjDE rank of 49, unranked in preseason AP poll, AdjTempo rank is close at 295)
Creighton – 0.5 (AdjDE rank of 40 is close)
Marquette – 3 (AdjDE rank of 154, 3PA/FGA rank of 62, unranked in preseason AP poll)
Kansas State – 2.5 (Luck rank of 282, unranked in preseason AP poll, AdjOE rank of 50 is close)
Wake Forest – 2 (AdjDE rank of 160, unranked in preseason AP poll)
South Carolina – (AdjOE rank of 149, unranked in preseason AP poll)
Miami (FL) – 3 (AdjOE rank of 68, unranked in preseason AP poll, AdjTempo rank of 338)
Minnesota – 2 (AdjOE rank of 81, unranked in preseason AP poll)
Vanderbilt – 3.5 (AdjDE rank of 41, 3PA/FGA rank of 6, unranked in preseason AP poll, AdjOE rank of 48 is close)
Dayton – 2.5 (AdjDE rank of 45, unranked in preseason AP poll, AdjOE rank is close at 49)
Rhode Island – 1 (AdjOE rank of 64)
Arkansas – 3 (AdjDE rank of 96, unranked in preseason AP poll, extreme caution required with Luck rank of 7)
Northwestern – 3.5 (AdjOE rank of 58, unranked in preseason AP poll, AdjTempo of 300, AdjDE rank of 35 is close)
Xavier – 1 (AdjDE rank of 74)
Michigan State – 1 (AdjOE rank of 66)
Virginia Tech – 3.5 (AdjDE rank of 125, 3PA/FGA rank of 90, unranked in preseason AP poll, caution required with Luck rank of 40)
Maryland – 2.5 (AdjDE rank of 64, 3PA/FGA rank of 72, caution required with Luck rank of 36)
VCU – 2.5 (AdjOE rank of 72, unranked in preseason AP poll, AdjDE rank of 36 is close)
Seton Hall – 3 (AdjOE rank of 76, unranked in preseason AP poll, AdjDE rank of 38 is close, caution required with Luck rank of 28)
Providence – 2 (AdjOE rank of 108, unranked in preseason AP poll)
USC – 3 (AdjDE rank of 86, unranked in preseason AP poll, caution required with Luck rank of 20, AdjOE rank of 46 is close)
According to this criteria, the best Final Four bets (per region) are as follows: Gonzaga, North Carolina, Louisville, and Duke.
Possible Final Four sleepers (among teams seeded 4th or lower): Purdue, West Virginia, Wisconsin, Wichita State, Creighton, Rhode Island, Xavier, and Michigan State. Creighton, Michigan State and Xavier do require an asterisk because of injuries they have sustained.

 

CHAMPIONSHIP CRITERIA

We are operating with the same parameters as the Final Four criteria here. As such, I recommend ensuring that all of your championship selections meet either all or all but one of the following criteria. Subjective accounting is still encouraged when evaluating teams on these criteria.

CRITERIA:

– EM rank of at least 6 or greater (12 out of 15 have met)

– AdjOE rank of at least 20 or greater (13 out of 15 have met)

– AdjDE rank of at least 40 or greater (15 out of 15 have met)

– Luck rank of at least 199 or greater (14 out of 15 have met)

– 3PA/FGA rank of at least 100 or lower (14 out of 15 have met)

– WS of 10 or less (15 out of 15 have met)

– Ranked in top 20 of preseason AP poll (12 out of 15 have met)

– Win at least one title – regular season or conference tournament (13 out of 15 have met)

The historical trends:

  • 12 of the past 15 champions (80%) have met all or all but one of the criteria; 8 (53.3%) have satisfied all 8 requirements.
  • The three that did not meet two or more criteria were 2003 Syracuse, 2011 UConn and 2014 UConn (see “Outliers” for more information).

Listed below are all championship contenders and the amount of criteria they do NOT meet. As such, 0 is excellent, 1 is passable, and 2 or more is problematic.

CHAMPIONSHIP CONTENDERS (3/12/17):

Gonzaga – 0
Villanova – 2 (3PA/FGA rank of 25, AdjTempo rank of 323)
North Carolina – 0
Kentucky – 1 (WS of 11)
West Virginia – 2 (Luck rank of 290, didn’t win a title)
Louisville – 2 (AdjOE rank of 23, didn’t win a title)
Florida – 5 (EM rank of 9, Luck rank of 292, unranked in preseason AP poll, AdjOE rank of 31, didn’t win a title)
Kansas – 1.5 (Be cautious with a Luck rank of 25, EM rank of 10)
SMU – 4 (EM rank of 11, AdjTempo rank of 330, unranked in preseason AP, WS of 16)
Duke – 1.5 (AdjDE rank is close at 39, EM rank of 12)
Purdue – 2 (AdjOE rank of 24, EM rank of 15)
Oregon – 2 (AdjOE rank of 19 is close, 3PA/FGA rank of 106 is close, EM rank of 16)
UCLA – 3 (AdjDE rank of 78, EM of 18, didn’t win a title)
Arizona – 1.5 (EM rank of 20, be cautious with a Luck rank of 37)
According to this criteria, the best bets for champion are as follows: Gonzaga, North Carolina, and Kentucky.

 

OUTLIERS

Over the course of the past 15 NCAA tournaments – and, indeed, in myriad years prior – numerous “Cinderellas” have emerged from the fray to make the Final Four when the popular opinion had them discounted or otherwise ignored. Although the performance of these underdogs flew in the face of convention, did it also fly in the face of our criteria? Let’s examine the 10 teams that did NOT meet 2 or more of the aforementioned Final Four criteria:

FINAL FOUR OUTLIERS:

2016 Syracuse (10 seed) – 5

2011 VCU (11 seed) – 4

2006 George Mason (11 seed) – 2

2011 Butler (8 seed) – 2

2010 Butler (5 seed) – 2

2014 Wisconsin (2 seed) – 2

2014 Florida (1 seed) – 2

2013 Wichita State (9 seed) – 2

2002 Indiana (5 seed) – 2

2015 Wisconsin (1 seed) – 2

A few things stand out immediately: first and foremost, 7 of the 10 outliers were seeded 5th or lower. I am of the opinion that there is nothing predictive to be gleaned from teams seeded that low making the Final Four; as discussed in the “Seeds” section, while we know that there is a good chance there may be a non 1- or 2-seed in the Final Four in a given year, our odds of choosing the correct one are slim enough to warrant a disregarding of the very exercise. It’s probably going to happen; we probably cannot reasonably predict it.

A second takeaway from the data provided by the outliers is that, despite the admittedly improbable nature of 5-, 8-, 9-, and 11-seeds all making the Final Four, all of those teams only missed out on 2 of the 7 criteria. Thus, although their appearance in the Final Four was unlikely, it generally adhered to the strictures and conventions established by the criteria.

Occasionally, the metrics at our disposal have tabbed a team not seeded on the 1 or 2 line as a legitimate Final Four contender – this can normally be attributed to the team in question possessing an Efficiency Margin ranked anywhere from 1st to 6th. Depending on the quality of the other highly seeded teams in their region, these teams can serve as exceptions to the notion that only 1- and 2-seeds should be selected for the Final Four. One poignant example of this occurred in 2006, as Florida, a 3-seed in the Minneapolis region, made the Final Four (and eventually won the title) over 1-seeded Villanova and 2-seeded Ohio State. Before the tournament began, both Villanova and Ohio State did not meet 2 or more of the Final Four criteria, while Florida only missed on 1. Additionally, Florida entered the NCAAs with an EM rank of 6th, only two spots behind Villanova and four spots ahead of Ohio State. Therefore, in this particular instance, choosing the 3-seed as a Final Four team was a lower-risk proposition than it would normally have been; this pick would have been completely justifiable.

Along the same lines, there have been 3 out of 15 champions that did NOT meet 2 or more of the championship criteria:

CHAMPIONSHIP OUTLIERS:

2014 UConn (7 seed) – 3

2011 UConn (3 seed) – 3

2003 Syracuse (3 seed) – 2

2014 UConn stands as the most improbable NCAA tournament champion of the 21st century for a variety of reasons. They were a 7 seed – the only seed lower than 3 to win in the span of this study. They entered the tournament with an Efficiency Margin rank of 25 – the lowest of any champion. Their Adjusted Offensive Efficiency was ranked 58th – the lowest of any champion. There was no logical reason why this team should win the title and, yet, they did. Inevitably, there will be another champion like 2014 UConn, one that defies all convention and wins despite the overwhelming mountain of statistics and history against them. My takeaway with regard to this is that there is no takeaway. As some say, “March happens”.

ADVICE: Ignore the outliers when choosing your Final Four and championship contenders. While they make for a romantic yarn on the pageantry and wonder of the NCAA tournament, determining who they will be is an unpredictable and inherently risky endeavor. If your goal is to pick the highest number of correct Final Four teams, stick with the numbers.

 

MISCELLANEOUS

For the tidbits that did not quite fit into any of the above sections …

On 3PA/FGA: There may be something to monitor moving forward with regard to the increase in 3PA/FGA across the entire sport of basketball and its correlation to NCAA tournament success. Historically, having 3 teams make the Final Four who finished the season in the top 60 of 3PA/FGA as we did in 2016 is an incredible outlier. If it happens again in 2017, we may have the start of a trend. TBD.

On FT% and NBA players: Two data points that I considered but ultimately excluded from this study were teams’ overall FT% and the amount of NBA players (and/or “talent”) that they had on their rosters. I recommend you evaluate them through the framework of a low team FT% and a lack of NBA talent being red flags in the pursuit of tournament success.

If you made it this far, thank you for indulging my research and me. This could not have been compiled without the work of Ken Pomeroy, Ed Feng and Shane Ryan – my sincerest thank you to all of those individuals. I will update the list of Final Four/championship contenders and their standing in various categories every Monday, culminating with the day after Selection Sunday. My hope is that this study is incomplete and will evolve over time.

– Matt Scharboneau

[1] Ken Pomeroy, “The Preseason AP Poll is Great,” http://kenpom.com/blog/the-preseason-ap-poll-is-great/, (November 1, 2010).

Contact Matt Scharboneau via email (mscharboneau12@yahoo.com).

Advertisements

7 thoughts on “Madness Explained: A Comprehensive Overview of Why Teams Make the Division I Men’s Basketball Final Four

  1. EastCoastHusky

    Thanks so much for doing all this research. Really fascinating stuff. I have a quick question for you- Purdue is listed as a team that meets all of the final four criteria, but they are left off of the Championship Contender list. What is the reason for that?

    Like

    Reply
    1. Matt Scharboneau Post author

      Thank you for reading! The Boilermakers were omitted because they did not project to be a particularly high seed and they failed to meet two of the criteria (their AdjOE rank is 24th and their EM rank is 15th). However, I still absolutely think they could make the Final Four, especially given that Kansas is a relatively weak 1-seed.

      Like

      Reply
      1. EastCoastHusky

        Thanks so much for the reply. I am a math teacher and I can’t tell you enough how much I appreciate this numbers based analysis. Will lean on this heavily as I fill out a bracket this week!

        Like

  2. James

    Great article! Question: Do you think Arizona’s EM increased after Trier returned? Is there anyway to calculate their EM going forward after he returned, and do you think it even matters?

    Thanks!

    Like

    Reply
    1. Matt Scharboneau Post author

      Thanks for reading! Interestingly enough, Arizona’s EM has actually decreased 8 spots, from 12th to 20th, since Trier returned. This is mainly attributable to the 85-58 defeat they suffered at the hands of Oregon – that game alone dropped Arizona from 14th to 22nd, and they’ve never fully recovered analytically. If you’re willing to completely disregard that result, Arizona would likely have an EM on the fringes of the top 15 – good enough to make the Final Four, certainly, but still serving as a red flag with regard to winning the title.

      There are a couple of additional problems with Arizona. Until very recently, their AdjTempo rank was in the 300s (something only 4 of 60 Final Four teams in this study have managed – it does help that they’re all the way up to 277 now, though), and their Luck rank is 37 (not impossibly high, but also not a rank Final Four contenders want to be at entering the tournament). It’s a shame, too, because either one of the Elite 8 teams that ran into Wisconsin a few years back would probably be a fine choice in this region.

      Like

      Reply
  3. larman@wowway.com

    I’m concerned with no consideration given to Strength of Schedule of all the clubs , but I do realize numbers do not lie …. Any thoughts on teams like Gonzaga ….I believe no club has reached the final 4 with their Strength of Schedule …. Thank you !!

    Like

    Reply

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