How to Pick a Winning Bracket

After the college football bowl season, I typed up a list of lessons that I learned from bowls in hopes that I would actually remember them come next year. The logic behind them was fairly sound, although I am quite certain that I will make the same mistakes again.

I thought about doing the same for March Madness this season, but I learned one lesson and one lesson only: unless you are in a pool with ten people or less, pick a unique team to win the tournament. That seems fairly simple…but it’s actually the opposite of how most people pick their brackets. We’ll get to that a bit later though. First, you get to read a story.

Every year I enter into a March Madness pool that has upwards of 500 people in it. I have never won it (and won’t this year either). The closest I came was in 2006, otherwise known as the year of George Mason. I couldn’t fully enjoy the Patriots’ run though, since they knocked off Connecticut in the regional final.

That year I went with the traditional formula for bracket selection. I picked several upsets in the first few rounds but picked a favorite to win the tournament. Almost all of my upset picks hit. I literally nailed everything. I hit seven of eight Elite Eight picks, only missing the aforementioned George Mason, which virtually no one picked to even win a game. My Final Four consisted of LSU, UCLA, Florida, and Connecticut. The first three all got there.

Then UConn happened. I picked them to win the tournament that year, but didn’t even need them to win. There were about 430 people in the pool, but I was so far ahead that all I needed was UConn to beat George Mason to lock up the title. I had the Huskies beating UCLA in the final. Because of that, no one that picked UCLA to win the title could pass me and no one that picked Florida to win the title could catch me if I got the seven points for UConn’s supposedly inevitable Elite Eight victory (and no one picked LSU to reach the title game at all).

Imagine – the $1200 or so first place prize locked up before the Final Four even began. We all know that the sports gods are cruel though. The Huskies blew it in overtime, I received no more points, and the handful of people that picked Florida to win the next two games (11 points for the semifinal and 16 points for the final) passed me. I finished in sixth place, five points behind the overall winner.

You would think the epiphany would have happened then. I really liked UCLA and Florida, but instead of picking either one of them to win the title, I went with the consensus favorite Connecticut. I would imagine that 30% of the people picked UConn to win it that year. I followed all of them and where did it get me? Sixth place. I think I won something like $50.

And for what? The sixteen points obviously wouldn’t have done me a whole lot of good, since everybody else picked them. Sure, had the Huskies won it all, I would have needed to pick them at least to the finals to hold the rest of the pack that did pick them. But I should have just taken my chances. Although I finished sixth, I ruined an amazing year of picks.

Had I picked Florida over UConn in the Final Four, I would have won something like $1500 between two pools. Had I done so and UConn actually did win, I would have probably won $0. I think I did win about $75 between the pools, but I would gladly (duh!) exchange that for picking Florida, a team I really liked.

Fast forward to this year. Finally, I went with an off-the-wall team to win the tournament. This time, I went with Kentucky almost completely by accident. I liked the Wildcats to do well in the tournament because they really came together at the end of the season, capped off by the dominant SEC tournament run. I didn’t like number one seeds Ohio State and Duke on that side of the bracket, so it was wide open. I also kept hearing about all the parity this year, so I figured why not take my chances with a #4 seed that is playing well.

Eventually I picked Kentucky to make it to the final, where they would lose to Kansas. Then I remembered my track record with Kansas. The last two times I picked them (2006 and 2010), they went out on the first weekend. I quickly erased Kansas and had them go down in the Sweet Sixteen. I ended up with Pitt in the title game for reasons that remain unclear to me. I don’t actually like Pitt, so I picked Kentucky to beat them in the title game.

Other than that, this was my worst year ever. That is not an exaggeration. My ENTIRE Elite Eight on the right side of the bracket was done by Saturday night on the first weekend. When the dust settled, Kentucky was the only team that made the Elite Eight that I got right. To put that in perspective, only four of the 490 people in my pool missed all Elite Eight teams…and you know that includes at least 40 women that picked based on name alone (my wife was one of them, and she performed far better than me). Of 490 people, I finished better than four people on Elite Eight picks.

The early rounds weren’t much better. I picked Florida State to reach the Sweet Sixteen in an upset pick. I picked Richmond to beat Vandy. And that’s it for picks that panned out. Michigan State and Louisville, both Elite Eight picks, went down in the first round. Pitt and Purdue, both Final Four picks, went down in the second. After the first weekend of play, I sat in 423^rd place. After the second, I sit in 348^th place.

But it turns out almost none of that matters, thanks to Kentucky. If Kentucky ends up winning the title, I will finish in fourth place. I only picked one Elite Eight team correctly, and I still have a legitimate shot to finish fourth in a 490 person pool.

All that time spent looking at first and second round upsets got me nowhere. Sure, I can be proud of my Seminole Sweet Sixteen pick. That got me a whopping three points. I would have had to do that with at least three more Sweet Sixteen picks to finish in first. The three people that would finish ahead of me are the only other three people that picked Kentucky to win.

So that’s my lesson. If you are in a pool with ten people or less, by all means follow the standard route. Pick a favorite to win and try to pick off enough points in the early rounds to win the pool. It is fairly unlikely that anybody will have any team lower than a two seed winning the tournament anyway.

But if you are in a big pool, like I am, you HAVE to go with an underdog to win the title. Don’t go crazy – no team lower than a four seed has won since #6 Kansas in 1988. And of course your #3 or #4 seeded pick actually winning the tournament is still an extreme longshot.

Yet the one strategic question I missed was this: what are the odds that I do well enough in the sixty pre-tournament games that I build up enough of a lead to top the 100 or so other people in the pool that picked the same favorite I did to win the tournament? Turns out quite a bit less than picking a significant underdog to win the title.

I understand that if UConn would have won just one more game in 2006, I would have still won the pool while picking the runaway favorite to win. That took a great deal of luck and magic to get me in that spot, and I seriously doubt I will end up in that position again. The larger point remains: even the biggest favorite to win the tournament will have less than a twenty percent chance to win. Why not roll the dice on a unique team to win the tourney and take my chances?

The lesson is simple yet a bit counterintuitive. I have picked NCAA brackets every year since I was a four-year old in 1989 (that is not an exaggeration). In my best ever year, I picked the runaway favorite to win the title and finished in sixth place. In my worst ever year, I picked an underdog to win the title and have a very realistic shot at finishing fourth place.

I like those odds. And go Kentucky!

Performances of Mid-Major At-Large Teams in the NCAA Tournament

I am endlessly fascinated with the performances of teams in the NCAA Basketball Tournament. This isn’t too surprising – I’ve already written about how I came to love sports by spending hours staring at statistics in the sports section of my local newspaper.

March Madness is the only popular event in which you hear about things like bubble teams, RPI rankings, and the like. So it makes perfect sense that I just spent hours researching how mid-major at-large teams have performed against comparable major conference at-large teams. I used to do stuff like this all the time (what did I do before the Internets!). Now that I have a blog, I figured that I might as well share it with the rest of my readers.

So here are the mid-major at-large selections since 2004 (because it was becoming too much of a pain to track down at-large teams from years earlier than that) and how they performed compared to teams that shared their same seeding:

2004 –
#1 St. Joseph’s – 3–1. Other #1 seeds went 4–1 (Duke) and 1–1 (Kentucky and Stanford).
#7 Memphis – 1–1. Other #7 seeds went 3–1 (tournament champion Xavier) and 0–1 (Michigan State).
#7 DePaul – 1–1.
#9 Charlotte – 0–1. The other #9 seed went 0–1 (Arizona).
#9 UAB – 2–1.
#9 Southern Illinois – 0–1.
#10 Louisville – 0–1. Other #10 seeds went 2–1 (tournament champ Nevada) and 0–1 (South Carolina)
#10 Dayton – 0–1.
#11 Richmond – 0–1. Other #11 seeds each went 0–1 (tournament champs Western Michigan and Utah).
#11 Air Force – 0–1.
#12 BYU – 0–1. Other #12 seeds went 1–1 (tournament champs Manhattan and Pacific) and 0–1 (tournament champion Murray State)
#13 UTEP – 0–1. Other #13 seeds all went 0–1 (tournament champs VCU, UIC, and East Tennessee State)

A huge year for mid-major at-large teams – twelve in all. I can only imagine that Doug Gottlieb and Billy Packer had aneurysms. Unfortunately we don’t have much to compare the teams to. Other than #1 seed St. Joseph’s, the best at-large team was UAB at 2–1, and we don’t have much to compare them to, since two of the other three #9 seeds were also mid-major at-large teams.

Two interesting thoughts when I pulled this up. First, I forgot how relevant Conference USA was in basketball before the Big East raided them for their basketball teams – a whopping six C-USA teams (Memphis, UAB, Charlotte, Louisville, Cincinnati, and DePaul) made the NCAA Tournament. Yet reading about this, all I can think about is how badly the Big East whiffed with not selecting Memphis. Instead they went with DePaul for the lucrative Chicago market (I’m sure that worked out well) and South Florida (they were good for that one half season of football, I guess).

Second, the last at-large team was 13th seeded UTEP. They were seeded BEHIND tournament champions Murray State, Manhattan, and Pacific. When the NCAA powers-that-be were thinking about expanding the tournament to 96 teams, they should have been forced to stare at that bracket for hours.

2005 –
#6 Utah – 2–1. Other #6 seeds went 3–1 (Wisconsin), 2–1 (Texas Tech), and 0–1 (LSU)
#7 Cincinnati – 1–1. The other #7 seed went 3–1 (West Virginia)
#7 Southern Illinois – 1–1
#7 Charlotte – 0–1
#8 Pacific – 1–1. Other #8 seeds went 0–1 (Minnesota, Stanford, and Texas)
#9 Nevada – 1–1. Other #9 seeds went 1–1 (Iowa St. and Mississippi St.) and 0–1 (Pitt)
#10 St. Mary’s – 0–1. Other #10 seeds went 2-1 (NC State) and 0–1 (tournament champ Creighton and Iowa)
#11 UAB – 1–1. Other #11 seeds went 0–1 (tournament champ UTEP and UCLA)
#11 Northern Iowa – 0–1

Nine mid-major at-larges were selected this year in what was certainly a backlash against the proliferation of mid-majors the previous year. The lower number of mid-major at-larges actually gives us a few more comparisons. Score one for Pacific against the major conference #8 seeds, but the three #7 seeds couldn’t combine for the number of wins that West Virginia had that year. That might not be exactly fair though, since the Mountaineers’ run spawned the phrase “What Would Mike Gansey Do?” that has been going strong for six years in the in-laws household. Hopefully I just sent you to Wikipedia to figure that one out.

2006 –
#7 Wichita State – 2–1. Other #7 seeds went 2–1 (Georgetown) and 0–1 (Marquette and Cal)
#8 George Washington – 1–1. Other #8 seeds went 1–1 (Kentucky and Arizona) and 0–1 (Arkansas)
#10 UAB – 0–1. Other #10 seeds went 1–1 (Alabama and NC State)
#10 Northern Iowa – 0–1.
#11 George Mason – 4–1. Other #11 seeds went 1–1 (tournament champ Milwaukee) and 0–1 (tournament champs San Diego State and Southern Illinois)
#12 Utah State – 0–1. Other #12 seeds went 1–1 (tournament champ Montana and Texas A&M) and 0–1 (tournament champ Kent State)
#13 Air Force – 0–1. Other #13 seeds went 0–1 (tournament champs Iona and Pacific)
#13 Bradley – 2–1.

Ironic that this tournament is remembered as the year of the mid-majors…but they were hosed pretty bad by the selection committee. Only eight berths, six of which made them an underdog based on seeding. And one of those (UAB at #10) was forced to play another underrated at-large team (#7 Wichita State) in the first round.

Led by Final Four participant George Mason, the mid-major at large teams went 9–8, despite not receiving a single seed higher than #8. Strangely, the other #11 seeds were all tournament champions. It seems that this practice has gone out of favor in recent years. Maybe it’s just me, but I don’t see a whole lot of mid-major tournament champions getting seeds higher than other at-large teams recently. Could be a coincidence or a small sample size though.

2007 –
#4 Southern Illinois – 2–1. Other #4 seeds all went 1–1 (Texas, Virginia, and Maryland)
#5 Butler – 2–1. Other #5 seeds went 2–1 (Tennessee and USC) and 1–1 (Virginia Tech)
#7 Nevada – 1–1. Other #7 seeds went 2–1 (tournament champ UNLV) and 1–1 (Indiana and Boston College)
#8 BYU – 1–1. Other #8 seeds went 1–1 (Kentucky) and 0–1 (Marquette and Arizona)
#9 Xavier – 1–1. Other #9 seeds went 1–1 (Michigan State and Purdue) and 0–1 (Villanova)
#12 Old Dominion – Other #12 seeds all went 0–1 (Illinois, Arkansas, and tournament champ Long Beach State)

And now we see the backlash from George Mason’s run to the Final Four in 2006. Had Butler and Xavier won their respective conference tournaments, we could have seen as little as four mid-major at-large selections. Not only that, but they matched up Butler/ODU and BYU/Xavier against each other. Tough to defend these selections.

Even tougher when you consider that every mid-major at-large team performed at least as well as their similarly seeded counterparts. Southern Illinois was the only #4 seed to even make the Sweet Sixteen this year. Butler was one of three #5 seeds to make it to the Sweet Sixteen, but none advanced further. Only fellow mid-major UNLV outperformed Nevada at the #7 seed, and the Rebels were assured of an at-large bid even without winning their conference tournament. BYU and Xavier canceled each other out in the #8/#9 game, but Xavier actually took eventual finalist Ohio State to overtime in the second round.

Finally, Old Dominion joined the other #12 seeds in the rare tournament without a 12 over a 5 upset. I distinctly recall picking three 12s to win this year – ODU, Illinois, and Long Beach State. I always tragically underrate Butler (I picked them to lose in the first round each of the last two seasons); I thought Virginia Tech sucked (they did, but Illinois sucked more); and Long Beach State gave up a ridiculous 121 points to Tennessee (which doesn’t even seem possible in college basketball – that’s more than three points per minute).

2008 –
#3 Xavier – 3–1. Other #3 seeds went 3–1 (Louisville) and 2–1 (Wisconsin and Stanford)
#7 Gonzaga – 0–1. Other #7 seeds went 2–1 (West Virginia) and 1–1 (tournament champ Butler and Miami (FL)
#8 BYU – 0–1. Other #8 seeds went 1–1 (Mississippi State and tournament champ UNLV) and 0–1 (Indiana)
#10 South Alabama – 0–1. Other #10 seeds went 3–1 (tournament champ Davidson) and 0–1 (Arizona)
#10 St. Mary’s – 0–1.
#11 St. Joseph’s – 0–1. Other #11 seeds went 1–1 (Kansas State) and 0–1 (Kentucky and Baylor)

Another swing and a miss for mid-majors. Only six teams got in for the second year in a row. The selection committee continued to stick mid-majors against each other. Including tournament champions from mid-major conferences, the first round saw Drake/Western Kentucky, Gonzaga/Davidson, UNLV/Kent State, and Butler/South Alabama. Maybe this was the fair way to do things…but the NCAA Tournament is a fan spectacle and the fun comes from Cinderellas beating Giants, not Cinderellas beating Cinderellas.

Xavier is actually the biggest mid-major supporter of them all: although they have won five consecutive Atlantic Ten championships, they have yet to win a single conference tournament during that time. In every year but 2010, the winner of the A-10 tournament would not otherwise have gotten in.

This was another huge year for mid-majors, it’s just that several of the teams that made deep runs qualified automatically. Stephen Curry’s Davidson squad made a dramatic run to the Elite Eight, Western Kentucky made it to the Sweet Sixteen, and #13 seeds Siena and San Diego each won their first round games.

2009 –
#4 Xavier – 2–1. Other #4 seeds were 2–1 (tournament champ Gonzaga), 1–1 (Washington), and 0–1 (Wake Forest)
#8 BYU – 0–1. Other #8 seeds went 1–1 (LSU and Oklahoma State) and 0–1 (Ohio State)
#9 Butler – 0–1. Other #9 seeds went 1–1 (tournament champ Siena and Texas A&M) and 0–1 (Tennessee)
#11 Dayton – 1–1. Other #11 seeds all went 0–1 (tournament champs VCU, Temple, and Utah State)

Yikes. Rock bottom for mid-majors. Had Butler and Xavier or Dayton won their conference tournaments, we could have seen as few as two mid-major at-large selections. Of course, part of the problem in the first place was that so many solid mid-major teams actually did win their conference tournaments – Memphis, Siena, Gonzaga, and Utah were all locks and Utah State and VCU were bubble teams.

In completely related news, this was the most boring tournament in recent memory. Each of the top four seeds advanced to the Sweet Sixteen in the East and South Regions. In the West and Midwest Regions, only #5 Purdue and #12 Arizona made it to the Sweet Sixteen. I don’t know what the cause and effect relationship was here. Maybe the NCAA was just that top-heavy that year. But I can’t help but think that the lack of mid-majors might have done fans a disservice.

The Final Four did turn out pretty appetizing this year: Michigan State, Connecticut, Villanova, and eventual champion North Carolina. It was just lacking in the middle section. A perfect tournament has three or four Cinderellas make it to the Sweet Sixteen and then the exciting powerhouses make it to the Final Four. This tournament only had the second part.

2010 –
#3 New Mexico – 1–1. Other #3 seeds went 3–1 (Baylor), 1–1 (Pitt), and 0–1 (Georgetown)
#7 BYU – 1–1. Other #7 seeds went 0–1 (Oklahoma State and Clemson)
#7 Richmond – 0–1.
#8 Gonzaga – 1–1. Other #8 seeds went 1–1 (Cal) and 0–1 (Texas)
#8 UNLV – 0–1.
#12 UTEP – 0–1. Other #12 seeds went 2–1 (tournament champ Cornell) and 0–1 (tournament champ New Mexico State)
#12 Utah State – 0–1.

We learned two things from this year’s tournament. First, we learned that the selection committee likes to over-compensate. The number of at-larges (seven) isn’t out of sync with past years, but the seedings certainly are. New Mexico was waaayy overseeded as a #3, #7 Richmond probably should have been a #9 or #10, and #12 UTEP was a strange selection to make the tournament at all.

The second thing that we learned is that mid-majors can do some serious, serious damage if given the opportunity. In many ways, this was the best year for mid-majors ever. Not counting the scandal-plagued Memphis program’s runner-up finish in 2008, Butler became the first real mid-major team to have a legitimate shot at a national title in decades. Northern Iowa, Cornell, and St. Mary’s joined the Bulldogs in the Sweet Sixteen. And Gonzaga, Murray State, BYU, Old Dominion, Ohio, and New Mexico all won games. A ridiculous eleven conferences were represented in the Sweet Sixteen.

You would think that the Selection Committee would take note – the 2010 tournament was by far the most exciting in recent history. Not only that, but this tournament almost single-handedly allowed the NCAA to expand to 68 teams and add three more money-making games to their slate. I don’t remember anyone ever talking about expansion until the 2010 tourney happened. Seriously, look back at 2009; there is no way they could have expanded after that season without a ridiculous uproar.

2011 –
#3 BYU – 2–1. Other #3 seeds are 3–0 (UConn) and 1–1 (Syracuse and Purdue)
#6 Xavier – 0–1. Other #6 seeds went 1-1 (Cincinnati) and 0–1 (Georgetown and St. John’s)
#7 Temple – 1–1. Other #7 seeds went 1–1 (Washington and UCLA) and 0–1 (Texas A&M)
#8 George Mason – 1–1. Other #8 seeds are 2–0 and 1–1 (Michigan)
#8 UNLV – 0–1.
#11 VCU – 3–0 and still alive. Other #11 teams are 2–0 (Marquette), 1–1 (Gonzaga) and 0–1 (Missouri)
#12 UAB – 0–1. Other #12 teams went 2–0 and counting (tournament champ Richmond), 1–1 (Clemson), and 0–1 (tournament winners Memphis and Utah State)

Lost in the shuffle of the outrage over UAB and VCU’s selections and the runs that mid-majors have gone on this year is the fact that the at-large squads kinda got screwed again. I have written about this in earlier posts, so no real need to recap it here, but it is pretty clear that we are doing something wrong with our selections. I mean, it’s not just that VCU was better than USC, Georgetown, and Purdue, it’s that they are FAR better.

I know I’m keeping my fingers crossed for a breakthrough performance for the rest of the tournament. And that’s what I mean by the fun of the tournament – you gotta believe!