Brackets and March Madness. It’s the perfect pairing but, ironically, it’s hard to get the pairing perfect.
The odds of picking every matchup correctly are one in 9.2 quintillions if you do a coin flip to make your picks. And if you make your picks based on your knowledge instead of chance? One in 120.2 billion.
In other words, picking the perfect bracket is impossible. However, picking a bracket that can beat out your friends or coworkers? That’s a different story.
Here are a few strategies you can use to fill out your bracket for optimal winning during March Madness in Oregon.
Use the seed-by-seed historical win percentages to pick upsets
Every year, there are upsets that bust brackets after the first day of games. But exactly how often do these upsets happen?
That’s a question the average basketball fan can answer: with simple percentages. Here’s a breakdown of the odds of each seed advancing to the second round:
- 9 over 8: 50.69%
- 10 over 7: 39.5%
- 11 over 6: 37.5%
- 12 over 5: 35.4%
- 13 over 4: 21.5%
- 14 over 3: 15.3%
- 15 over 2: 6.3%
- 16 over 1: 0.7%
Using this data, you can draw some pretty straightforward conclusions.
There are four of each seed in the tournament. If around 50% of 9 seeds and around 40% of 10 seeds win every year, on average, then you can pick two 9 seeds and one 10 seed with relative confidence. The same goes for your 11-6 and 12-5 matchups.
Once you get to the 13 seed, things are dicier. On average, fewer than one 13 seed each year wins a first-round matchup.
There have been five tourneys over the past 20 years in which no 13 seeds won. There have been four years in which two 13 seeds got a win. And there are 11 years in which one 13 seed has won.
You’re looking at roughly a 50% chance that one 13-seed wins, and a 20% chance that two 13 seeds win.
The story changes with the 14 seeds, though. Over the past 20 years, only nine of 80 teams with the #14 seed have won, which is around 11%. The odds drop even lower with the 15 and 16 seeds.
What does historical win percentage reveal? That you should avoid picking upsets by 14 seeds and higher and be very careful about choosing your 13 seeds.
Use the average upsets per tournament method
If you don’t want to roll the proverbial dice with percentages, you can take a simpler approach to build your bracket: look at the average number of upsets per round.
Here’s the average per-year data from the NCAA:
- Total upsets per tournament: 12.4
- First-round upsets: 6.2
- Second-round upsets: 3.7
- Sweet 16 upsets: 1.7
- Elite Eight upsets: 0.5
- Final Four upsets: 0.2
These numbers reveal that upsets represent around 10% of outcomes in the first three rounds, then drop off considerably in the Elite Eight and Final Four.
Using this data, you could simply select six upsets in the first round, four in the second, two in the Sweet Sixteen and none in the Elite Eight and Final Four.
Just remember that, because the NCAA tournament has been around for decades, there is some variation in the data. For example, there were only two first-round upsets in 2007 but 10 in 2016.
It may help to pair the average-upsets-per round metric with the per-seed winning percentages. For example, if you are going to pick six first-round upsets based on win percentage, you might choose two 9 seeds, then one team each from the 10, 11, 12, 13 seeds.
Use RPI to predict win totals for 15 teams
RPI, or “rating performance index,” is a metric the NCAA Selection Committee used to use when ranking teams for the tourney.
The metric has its detractors, but, in general, RPI can be a reliable way to predict the number of wins a team will get in the tourney. Here’s a breakdown of average wins per tourney based on RPI ranking from 2010 to 2017:
- #1: 2.75 wins
- #2: 2.5
- #3: 3.00
- #4: 3.25
- #5: 2.63
- #6: 3.38
- #7: 2.00
- #8: 2.38
- #9: 1.75
- #10: 1.13
- #11: 1.38
- #12: 1.00
- #13: 0.71
- #14: 2.88
- #15: 1.00
These numbers reveal some fascinating trends. First, the team ranked 13th in RPI averages less than one win a tournament, which makes them a prime candidate for a first-round upset.
The RPI #6 wins the most games, on average. Seed-wise, a sixth-ranked team in RPI would, in theory, be a #2 seed. And #2 seeds have a relatively easy path to the Sweet Sixteen since they have around a 40% chance of playing a #10 seed in the second round.
Focus on second-round upsets
If a bracket isn’t busted after the first round, there’s a good chance it can self-destruct in the second round.
To avoid that, look at the upset data. According to the NCAA, over the past 36 years, the six seed has beat the 3 seed 29 times. Here’s a rundown of the rest of the upset frequencies in the second round:
- 7 over 2: 26
- 10 over 2: 18
- 11 over 3: 18
- 8 over 1: 14
- 12 over 4: 13
- 9 over 1: 6
- 13 over 5: 3
- 14 over 6: 2
- 15 over 7: 2
Based on these numbers, the rules for the second round are:
- Pick at least one #2 seed to lose in the second round
- Never pick a 13 seed or higher to advance to the Sweet Sixteen
- If you’re going to pick a #1 seed to lose, choose an 8 seed to do the deed, not a 9.
Pick the high seeds to go to the Final Four and Championship
There have been 144 teams that made it to the Final Four since the tourney went to 64 teams in 1985. Of those 144 teams, 119 teams (around 83%) have been a 1-4 seed.
So, if you’re playing the odds, your Final Four should only be made up of 1, 2, 3, or 4 seeds.
If you want to pick an upset, your best chances lie with picking either an 11 or 5 seed, as they’ve accounted for around 10% of Final Four teams.
Who should your Final Four winners be? Here’s what the historical numbers reveal about the seeds that make it to the finals:
- 1 seed: 50% of teams in the final
- 2 seed: 25%
- 3 seed: 25%
- 4-8 seeds: 25%
The takeaway: Your championship game should have at least one #1 seed. After that, it’s a toss-up. As for picking your champion? History says the 1 seed is the favorite, as they’ve won 23 of the last 36 finals.
The 2 seed has won five titles, the 3 seed has won four, and seeds 4-8 have won four.