Sign inSign up free
Back to CBB

2021 March Madness Game Theory Optimal Bracket

March 16, 2021 · Mat Houchens · CBB

Bracket time! We missed out on the madness last year due to COVID but this year we finally have a bracket tournament. And if you're like me, you probably participate in at least one bracket contest. I've had a shot to win a few times over the years but haven't come away with the win yet but I trust the process.

The contest I play in has thousands of brackets. When you play in contests of this size, it's not just as simple as advancing favorites of each game. Maybe if your goal is to get as many correct games as possible, that's a potentially good strategy but your goal should be to try and win the bracket contest.

In most bracket contests, points double each round meaning you get more points for later games guessed correctly. Some of you are probably even in an upset pool where you get bonus points for seed differential. This is where being contrarian will work to your advantage going against the grain.

Hometown Pools

Know the bracket contest you're in. I'm from Gainesville so any time the Gators make the tournament, they are usually always over compensated and their pick % is very inflated. For example, in 2014 when they were a #1 seed they were picked by 36% of the 2,366 brackets to win it all. Absolutely egregious.

For the record, FiveThirtyEight had Florida chance of winning it all at the beginning of the tournament at 14%. That’s a huge surplus of confidence in my fellow Gainesville bracketeers.

There is 0 value in picking Florida betting with my heart in this instance. I would gladly donate $20 to a contest to have my favorite team win a championship.

But if I pick Florida to win it all in this example, I still have to out perform 36% of the field with the rest of my pics. That doesn't really excite me to get the champion right and still competing with over 1/3 the field and a good chance of still winning absolutely nothing.

The point I’m trying to make is know your pool. If there’s a hometown bias and that team makes the tournament, think twice before moving them too far regardless of the seed they received.

Odds to Win vs Pick %

This is probably the easiest way to go.  You can find some overall inefficiencies with the people's bracket on ESPN.

I tend to agree with this thought process and have been doing it this way for years. In most contests you get more points for later rounds so typically if you have your Final Four correct, you're in a great spot vs someone that somehow got all the random upsets right in the earlier rounds but missed out on having all four teams right.

Right now, Gonzaga has about 27% chance to win per FiveThirtyEight but are being picked around 36% of the time. Not only are you going against a huge % of the field, you're picking a team that the public is overconfident on. That's not a good feeling. You can immediately get rid of over 1/3 the field by not picking Gonzaga to win it all and even further separate by having them lose earlier.

Below is a table of the top 8 overall teams chance to win the tournament and their odds and pick % by round. Honestly any teams outside of the list, you're already decently contrarian.

  • Team
  • Elite 8 is the 538 odds to win in the Elite 8 round to go to the Final Four
  • ESPN 8 is the ESPN people's bracket pick %
  • Then the plus / minus of that decision at that point
  • Final Four is the 538 odds to win in the Final Four to go to the National Championship
  • ESPN 4 is the ESPN people's bracket pick %
  • Then the plus / minus of that decision at that point
  • NCG is the 538 odds to win the National Championship game
  • ESPN Winner is the ESPN people's bracket pick %
  • Then the plus / minus of that decision at that point

A positive % means an edge on the field whereas a negative % means you're going with an overconfident public. So for example, you can see putting Michigan to the final four looks to be really over confident but you can see that decreases greatly to put them in the championship game and even win it all you get some of that equity back. So what that means is if you put Michigan in the Final Four you can make up some equity by moving them on to the finals as the majority of the field is putting them in the Final Four and then having them lose to Gonzaga so you can make up some of that ground by having them beat Gonzaga in the Final Four but that's still likely to be a popular bracket decision overall and really depends on what you plan to do on the other side of the bracket.

Team Elite 8 ESPN 8  + / - Final Four ESPN 4  + / - NCG ESPN Winner  + / -
Gonzaga 54.5% 66.6% -12.1% 41.8% 52.9% -11.1% 27.2% 35.7% -8.5%
Illinois 44.9% 53.2% -8.3% 27.7% 31.4% -3.7% 15.4% 15.3% 0.1%
Baylor 40.5% 48.9% -8.4% 24.0% 28.1% -4.1% 13.2% 10.0% 3.2%
Iowa 17.8% 14.2% 3.6% 11.1% 7.8% 3.3% 5.6% 3.5% 2.1%
Houston 21.2% 16.2% 5.0% 10.8% 7.6% 3.2% 4.9% 2.8% 2.1%
Michigan 23.9% 40.1% -16.2% 9.0% 14.0% -5.0% 4.2% 8.0% -3.8%
Ohio State 19.1% 23.1% -4.0% 8.8% 9.4% -0.6% 3.8% 3.5% 0.3%
Alabama 20.2% 21.8% -1.6% 7.3% 6.2% 1.1% 3.3% 2.9% 0.4%

Here's a link to download the CSV if you want to do more work with it.

My 2021 March Madness Picks

Let me preface this again by saying I play in a large bracket contest and this shit is such a crap shoot trying to determine so I tend to let the numbers help me decide. I haven't watched much CBB this year so I will rely even heavier on zigging where the public is zagging (get it?)

Top Fade

  • Gonzaga - across the board, the public is over confident in all aspects of their odds. I really like having them lose to Iowa in the Elite 8 to differentiate my brackets quite a bit and helping me if I choose to move Michigan on past the Final Four.

Toss Up

  • Michigan - this is a tough one for me but I either am having Michigan go all the way to the big game with a Gonzaga team getting knocked out before the Final Four or having them lose to a Florida State or Alabama later.

Under the Radar(ish)

  • Baylor
  • Iowa
  • Houston
  • Ohio State

I really like putting Iowa in the Final Four as mentioned above to differentiate. I think a great option to further separate is having Baylor beat Illinois to face Michigan or Iowa in the championship game.  I don't think you have to get too crazy to help yourself separate from the pack given some of these teams odds to win it all like Baylor is 13% but only getting picked 10% of the time. Just make some key pivots and all of a sudden you find yourself with a unique chance to take down a big contest.

Final Four March Madness Picks

  • Iowa
  • Michigan
  • Baylor
  • Illinois

And then I think I lean towards Michigan beating Baylor for the title.

Good luck and may you run pure in your bracket!  What do you think? Who are you picking for your contest?