2024 Bracket Picking Guide
General Bracket Theory
There are a lot of articles out there to help you pick your bracket. Most of them will give you some statistics about past champions and a rule about not picking freshman point guards. I’m aiming to do something a little different and provide a more mathematically sound basis to bracket picking.
It may sound a bit silly, but the first thing you need to figure out when filling in your bracket is what exactly you are trying to accomplish. If you just want to maximize your expected score, the optimal strategy is quite easy- pick only favorites. You’re quite likely to end up above the median bracket with this strategy but quite unlikely to make it to the very top of the leaderboard.
More likely you want to win a pool ranging in size from 5 to 5,000 people. It’s important to note the size of the pool you’re trying to win- the bigger the pool, the more aggressive you need to be. A general rule of thumb I like to follow is that if you’re trying to win a bracket pool of n people, you want to maximize your chances of finishing in the top 1/n brackets. In plain English, if you want to win a 10 person pool, you want to maximize your odds of finishing in the top 1/10 of brackets. Similarly, if you want to win a 5,000 person pool, you want to maximize your odds of finishing in the top 1/5,000 of brackets. Thus, the bigger the pool, the more aggressive you need to be.
The key is to identify which teams are underpicked relative to their true likelihood of winning the tournament and pick them. Just how underpicked of a team you want to go for depends on how aggressive you need to be. If you’re trying to win a 5,000 person pool, you may need to really shoot for the stars, and pick a 1/500 longshot that only 1/5,000 people are picking in their brackets. If you’re trying to win a 5 person pool, you can be a bit more by the book. I generally like to go for an unconventional national champion and then relatively consensus Final Four picks elsewhere. This was my strategy last year and it worked brilliantly as I won my 80 person pool, picking UConn to win it all.
Without further ado, here’s some data on how frequently various teams are being picked (data courtesy of Yahoo!’s bracket challenge) as well as various metrics of how likely each team is to advance to each round.
2024 Data- Overall
Some brief column definitions:
Computer Avg.: The team’s probability of winning the tournament, averaged over the three computer systems shown.
Vig Adjusted Market: The team’s probability of winning the tournament, based on Vegas betting odds (odds courtesy of Circa).
Picked: What % of users are picking the team to win their bracket.
Diff: “Vig Adjusted Market” - “Picked”
Ratio: “Vig Adjusted Market” / “Picked”
The first thing that stands out here is that Connecticut is picked by nearly 1/3 of users, but by most metrics has around a 1/6 chance of winning the tournament. Despite being the most likely to win according to Vegas, there’s no value in the Huskies.
Depending on how big of a pool you’re trying to win, there’s varying degrees of aggression you can take. For a typical pool size of 5-25 brackets, Purdue, Arizona and Tennessee are all reasonable picks. Houston is beloved by the computers but strangely not by the betting markets (this was also the case in 2022). In a 50-100 person pool, Auburn is probably your best bet.
Once your pool size gets into the triple digits, you need to start looking further afield for a champion. BYU and St. Mary’s are decent options, and if you want a real wildcard, maybe take a shot on a New Mexico squad that is a top 30 outfit at most computers and has a favorable first and second round draw.
2024 Data- East Region
Connecticut is picked to make the Final Four by almost 2/3 of users but has only about a 40% chance of making it there. If you’re going to go with an unconventional Final Four pick somewhere, this could be the region to do it. Any of the next three teams- Iowa State, Illinois and Auburn all are underpicked.
2024 Data- South Region
The thing you don’t want to do here is pick Kentucky who is picked in over 20% of brackets but is outside the top 15 in every computer. Going chalk with Houston is a reasonable pick. I am pretty shocked that more people are not picking Duke given their brand, they’re surprisingly underpicked.
2024 Data- Midwest Region
This is probably the least clear region. Creighton could be worth a punt. Gonzaga would be a better pick if they hadn’t drawn an awful first round matchup with McNeese. Purdue and Tennessee are both being picked right in line with their fairs.
2024 Data- West Region
North Carolina is the weakest 1 seed by far, and traditionally picking the worst 1 seed to go down early is a nice strategy. Oddly enough- more people are picking the Tar Heels to make the Final Four than are picking Purdue or Houston. That’s the power of name recognition, I guess.
This is the most wide open region and I couldn’t fault you with picking any of the 2 through 5 seeds to advance, I kind of like a flyer on Alabama who has an easy first round matchup and a talented roster.
All in all, bracket picking is part art and part science. Hopefully after reading this you have a bit of a better idea as to what sorts of risks are worth taking this year and what sorts of risks are not. Best of luck!