Analyzing My Bracket Hunches

As we are in the grips of March Madness and filling out brackets I figured it would be a fun time to breakdown the “rules”, in some cases more like advanced hunches, I’ve developed over the last 20 years of making picks and see if they’re actually useful or actively harmful. All tournament performance stats below will be from the 2004-2024 tournaments.

My favorite teams go to at least the Sweet Sixteen

My obvious favorite team is Virginia Tech because that is the school I attended. Other schools I consider favorites that get preferential treatment as well include my childhood favorite Maryland, and Michigan and Michigan St. from familial ties. Georgetown would be on the list of favorites but they sadly cannot be trusted in March. If one of those schools are in the bracket they’re in the Sweet Sixteen for me and if I can justify it even a little bit logically they’re going to the Final Four. The Final Four is usually where I’ll have my preferential picks stall out, I believe the only time I picked my favorite team to win the title was the 2023 Virginia Tech women since they were a 1 seed.

Wow a combined 41 percent Sweet Sixteens! Do not look into who is being carried here! Don’t do it!!

Mid-Major local favorites win at least one game

This is related to number 1 but I reserve my most irrational bracket work for my Power 4 favorites. For my other favorite mid-majors from around Virginia (George Mason, VCU, ODU, Richmond), I occasionally pick them to the Sweet Sixteen or beyond but generally only reserve them for a first round win. I wish that I had put Norfolk State in this category in 2012 when they upset Missouri as a 15 seed, but sadly it was not to be.

For a group of teams that are often seeded 10 and below a 42 percent first round win rate is nothing to sneeze at. Advancing them to the Sweet Sixteen does not quite have as robust a statistical backing but you can’t fight your heart.

The teams I hate go no further than the Sweet Sixteen and lose as early as possible

This is a pretty easy list as it’s just Virginia, Duke and UNC. They stink and deserve to be sad.

Yes I am very happy to see the times upset number is bigger. No I do not care that it’s because it includes things like losing to a 4 seed as a 1 seed in the Sweet Sixteen or losing to an 8 seed in the Final Four.

Florida, Xavier, Oregon, Syracuse, and UConn always make it to the Sweet Sixteen

Over the years you notice certain programs seem to always be appearing in the Sweet Sixteen and you just know to automatically move them there regardless of seed or opponent. I don’t particularly have any affinity for these schools, but I know them to be reliable. Is this actually true? I don’t know I’ve never bothered to check until now.

The stats seem to back this faith up pretty well with a robust 56% Sweet Sixteen rate and 38% of the time they’re making the Sweet Sixteen despite having a seed lower than 4.

St. Mary’s, Illinois, Purdue, Houston, Georgetown and Tennessee cannot be trusted

Similar to the previous category but in the opposite direction, these are all mostly teams I’ve trusted repeatedly in the past only to get burned (except for Illinois because I always knew Brad Underwood stunk). This collection of teams often enters the tournament with good seeds and even better advanced metrics but it’s all a mirage because they will lose early to someone they shouldn’t.

I feel total vindication here. Despite frequent years with high seeds these teams cannot get past the Sweet Sixteen and are getting upset almost 1/3 of their appearances.

One of the First Four at-larges will make it to the Sweet Sixteen

Pretty self-explanatory. Starting with the first entrant in VCU along to fellow Final Four participant UCLA, it feels like every year one of these First Four at-larges will make it at least to the Sweet Sixteen if not further. No fancy table needed here the First Four has been around since 2011 and one of those at-large teams has made the Sweet Sixteen in 5 out of 13 years (38%). Overall, 12 out of 26 First Four at-large teams have won at least one game. Those are pretty good rates for what is usually an 11 or 12 seed.

Exactly one 8/9 seed will beat a 1 seed

This is common sense to me that is often not always actionable in a way that’s beneficial to bracket success. I am usually foolhardy enough to decide I can guess which one will be the one to lose and when it’s right (like Loyola Chicago over Illinois in 2021) it’s a great feeling. When it’s wrong it often ends up with that 1 seed in the Final and another disappointing showing in the bracket challenge. In the past 20 years at least one 1 seed has lost to an 8/9 seed in 10 times. Notably, since 2010, the only years that did not have a 1 seed lost to an 8/9 seed were 2012, 2016, 2019, and 2024. And it really only does happen once per year. The two years multiple 8/9 seeds made the Sweet Sixteen was when UMBC and Fairleigh Dickinson beat Virginia and Purdue.

No more than two one seeds in the Final Four

First, it’s just boring to pick chalk. As you can probably tell from the previous six rules I’m in this more for fun than statistical maximization. I think the glory comes from making an oddball pick that pays off more than following what the seeds tell you to do. Without looking it up two seems like the most frequent occurrence over the years, but ultimately I would lean more towards none than I would four.

My blind judgement seems mostly correct although it turns out years with only one 1 seed making the Final Four just edges out years with two. And look, there have been more years with zero one seeds in the Final Four than there have been with three and four combined. So let that be a sign to follow the hunches and have some fun out there!

If you have any similar hunches of your own I’d love to hear about them.