Wednesday, March 18, 2015

March Madness

Few could have imagined 30 years ago, when the field was first expanded into the 64 team bracket we know and love, just how big of a slice of Americana March Madness would become.  Every year tens of millions of brackets are produced, millions of TV guides are searched to remember where the heck TruTV is, and Sandy from accounting (you know, the one picking SMU to win because those are her initials too) wins the office pool with disconcerting frequency.


There are over 9.2 quintillion (9,200,000,000,000,000,000) possible bracket combinations in theory, so Warren Buffett was riding some really strong odds when he offered a mere billion dollars compared to this daunting number.  Of course not all matchups are created equal, and even the most basic of safe assumptions (a 16 has never beaten a 1 in the men’s bracket) makes 2^63 odds reduce down to a still-ridiculously-daunting 2^59. 

But while there’s still some time to log in to ESPN and change around a few games, I thought I’d offer you some facts, stats, and trends to help make your selections with a little more knowledge behind them.

The True 1 Seeds
Every year the selection committee makes its picks for the top 4 teams, and the track record on these is pretty strong overall.  In 30 tournaments, a 1-seed has won 18 times.  But an even more frequent predictor of success has been the Simple Rating System model (margin of victory weighted against a team’s strength of schedule, to explain it simply).  In 23 of 30 tournaments, a team with a top-4 SRS has won the title (9, 5, 5, and 4 times by individual rank 1-4). More recently, 5 of the last 10 champions had the best SRS ranking of the year, with 2 finishing 2nd and 1 in 4th.  Only 4 of 30 champions have won despite a double digit SRS ranking, though UConn’s 2 wins in the last 4 years came with rankings of 16 and 20. 

Entering the tournament this year, the top 4 teams in this metric are pretty much who you would expect: Kentucky (29.45), followed by Wisconsin (25.00), Arizona (24.80) and Duke (24.14).  It’s important to remember that SRS continues to adjust up or down for tournament games as well.  Five more teams are within 3 points of Duke’s current total, and could conceivably break into the top 4 with a title victory.  These are Villanova (23.33), Virginia (22.62), Gonzaga (21.90), then the less obvious duo of North Carolina (21.32) and Utah (21.30).  So while it wouldn't be unprecedented for another team to make a run, history strongly points to a champion emerging from this group of 9.

Avoid Pure Chalk
Every year the temptation to look and simply pick the 4 1-seeds to all make the Final Four is there, yet a Final Four without a 1-seed has happened more often than a year of all 1-seeds (twice vs once).  Of course a 1-seed’s odds are historically better than any other, but only 40% of Final Four teams in the last 30 years have come from the 1 line. 

So don’t be afraid to find one or two value picks from further down the bracket to mix into the Final Four equation.  There are great value picks of teams seemingly under-seeded at the 4 through 7 lines with N. Carolina, Utah, Butler, and Michigan St. leading the way on each one’s respective line.  Even Ohio St (SRS rank: 10) and Texas (SRS rank: 18) offer great value at the 10 and 11 line if you have enough faith that either the ‘stuck in neutral for 2 months’ Buckeyes or the ‘can’t close out games’ Longhorns have discovered a little magic in their prep time.  This year I've split the baby with 2 1-seeds and 2 non-1s in my Final Four.  My Final Four predictions to come at the end. 

The Unstoppable Force vs The Immovable Object
Rarely does the tournament offer us a showdown of the nation’s most efficient offense against its most efficient defense, but this year the potential for such a matchup is fairly high in the Midwest region.  Kentucky, as one could expect, gives up the fewest points per 100 possessions in the nation at a stingy 83.6.  Lurking as a credible Elite Eight opponent for the Wildcats is the 3-seed Notre Dame, top offense in the nation with a whopping 121.3 points scored per 100 possessions.  As intriguing as the matchup on that half of the court could be however, the Irish’s chances for the upset are dimmed greatly by their extremely mediocre defensive efficiency, 168th in the nation (Kentucky’s offense rates a still very strong 11th).

Riding the 12 seed
If you’re interested enough in the NCAA tournament to read a post like this, then surely you know of the consistent phenomenon of the 12-5 upset.  It has happened 44 times in the last 30 years, including 8 times out of 12 in the last 3 tournaments.  It’s an upset that makes a lot of sense when one thinks about it.  Five seeds often are either middling teams from power conferences (hello West Virginia), or good teams from conferences with suspect overall depth (Arkansas, Northern Iowa).  Twelve seeds on the other hand have typically been the very best of the champions from the 1 bid mid-major leagues, schools with a combo of talent and/or experience that set them apart from the other 18 or so schools in the mid-major genre. 

The obvious candidate of 2015 to continue this trend is Buffalo, champions of a Mid-American conference that grades out extremely well this season and a team that has played well in limited interaction with elite teams (a 5 point halftime lead against Kentucky standing out).  They face a West Virginia team that has some strong wins on their resume courtesy of the highly rated Big 12, but rather weak metrics and a simple lack of sustained momentum during the season.  Arkansas is a 5 with similar unremarkable numbers, though their opponent in Wofford lacks the strong numbers Buffalo can post.  Stephen F. Austin against any of the other 5-seeds would be a very strong contender, but their matchup against a Utah squad with numbers more akin to a high 3-seed than a 5 dampens my enthusiasm.

Something to note: if you’re picking 2 or more 12 over 5 upsets, odds say you should have 1 of these teams continue on into the Sweet 16.  12-seeds who win in round 1 are a relatively solid 20-24 in the round of 32, though their win percentage after that drops to single digits.  Buffalo is my leading contender for this role as well, as their follow up game would likely be against a Maryland team that is similarly unimpressive when you break down their numbers.

Don’t Fear The Big Upset
Many brackets avoid betting against any top-4 seeds to lose their opening games, yet the numbers say otherwise.  43 times a 13 or 14-seed has won in the first round, just 1 fewer win than 12 seeds boast.  Of course when you start getting this far down the seed line the metrics will never give you a clear choice to make (though Valparaiso is, by 13-seed standards, exceedingly strong), so playing a hunch on a team is the way to go.  4-seed Georgetown has elevated losing to double-digit seeds into an art form over the last several years, and their matchup against Eastern Washington sticks out like a sore thumb when trying to identify where this year’s biggest Cinderella could emerge (Georgetown’s metrics that identify it as a 6-seed masquerading as a 4 don’t hurt the case either).

Predictions
Well if you've slogged all the way through that then you at least have earned some predictions from me that you can mock ceaselessly if/when they fail.  So here they are

Cinderella Sweet 16 teams (7 seed or higher): (12) Buffalo, (7) Michigan St.
Final Four: (1) Kentucky, (2) Arizona, (1) Villanova, (5) Utah
Championship game: Arizona vs Villanova
National Champion: Arizona

Some explanations for each – No team in America matches up better to take down Kentucky than Arizona in my estimation.  Undervalued from the limited media exposure that often plagues teams out west, I like them to do to Kentucky what Duke did to the last unbeaten to reach the Final Four, UNLV, back in 1991.  They're also 1 of only 4 teams that boasts a top-20 offensive and defensive efficiency rating (Kentucky, Utah, and Wichita St. being the other three).  They have the depth needed to not tire out, with 10 players averaging at least 3 points and 7 minutes per game and 6 averaging over 9 points and 20 minutes per game. I like them to survive a tough Elite 8 matchup with Wisconsin before pulling the big upset of Kentucky and cap it off with a third win over a 1-seeded Villanova, similar to the school’s 1997 title run.

Utah is my shock pick for the Final Four.  Very underrated by the committee as a 5-seed, I like them as the team I have the fewest overall concerns about in a rather thin South region (watch out for 6-seed SMU as a potential Elite Eight team in the South too).  Villanova's path to Indy will be aided by a Michigan St. upset of 2nd seeded Virginia, a team that has lost its magic after battling injuries.  Kentucky looks good to roll through the Midwest, though I do have a bracket out there that continues my tradition of always picking Kansas in at least 1 bracket to win the title. 


Whatever picks you ultimately settle upon, let's not forget what makes the tournament the two best sporting weeks of the year.  No matter how big or small the school, every single team has the chance to go out and earn themselves a title on the court without some computer, pollster, or pundit blocking their way.  So armed with a lot of knowledge and a little luck, here's hoping your brackets are filled with circles not lines!

No comments:

Post a Comment