The approach outlined here is similar to the one I used last year to correctly identify six of the Elite Eight teams — San Diego State, Kansas State, Miami, Texas, Connecticut and Gonzaga, none of which were No. 1 seeds. That approach also helped me predict that fourth-seeded Connecticut would make the Final Four, and that fifth-seeded San Diego State would reach the title game, picks that helped plenty of my readers win their pools. If you followed the “People’s Bracket” last year — the most popular picks for every round in ESPN’s contest — you would have finished with 45 points in standard scoring systems, with only one Elite Eight team and none of the Final Four. My bracket finished with 96 points.
(If you have specific questions, please submit a note for our Reader Q&A. I’ll be answering questions Tuesday at 1 p.m. You can even upload your full bracket and let me critique it.)
To build our strategy, we’re going to incorporate historical trends, make some educated guesses based on analytics and lean on betting markets to point us in the right direction. Some of these tips may be new to you, but rest assured this is the correct path to take — and a far more productive blueprint than making picks based on school colors, mascots or the advice of noisy TV pundits.
Don’t start filling out your bracket with Round 1; start with the Elite Eight or Final Four
Most people sit down with their brackets, start at the 1 vs. 16 matchup in the top left corner and work their way through that region until they come to the Final Four. We’re not most people. We start with the teams we think will reach the Elite Eight or the Final Four and work backward to reduce the number of decisions we need to make.
Why? Because according to a March 2020 study in the Journal of Quantitative Analysis in Sports (“Models for generating NCAA men’s basketball tournament bracket pools”), bracket generators that start by selecting the teams that reach the Elite Eight or Final Four tend to outperform generators that start with the round of 64 or the national champion. It’s also the “best for balancing initial pick risk against the number of decisions,” per Sheldon H. Jacobson, a computer science professor at Illinois and one of the authors of the paper.
Check back later for my latest “Perfect Bracket,” but here’s a spoiler: No. 4 seed Auburn looks awfully enticing. The Tigers enter the tournament as the fourth-best team in the country, per analyst Ken Pomeroy’s ratings, which adjust margin of victory for tempo and strength of schedule. The Tigers also have one of the best defenses in the country. They’re in a powerhouse East Region with No. 1 seed Connecticut, No. 2 seed Iowa State and No. 3 seed Illinois, but that just means fewer of your competitors focus on them.
Look for value in the Elite Eight
According to data from ESPN’s past tournament pools, No. 1 and No. 2 seeds are typically well represented in the public’s Elite Eight. An average No. 1 seed is advanced to that round by 63 percent of entrants, while an average No. 2 seed is found in the Elite Eight on nearly half of all brackets. That tells us we should look elsewhere for our choices when possible, especially if we’re entering larger pools.
Specifically, look for highly rated teams in Pomeroy’s rankings that have been under-seeded in the tournament. You could also target lower-seeded teams that have a high consensus rating, using analyst Ken Massey’s aggregation of dozens of rating methods, relative to the field. In addition to Auburn, a team to consider is No. 5 Saint Mary’s in the West Region. The Gaels won the West Coast Conference regular season and tournament titles, reeled off 16 straight wins at one point this year and rank 20th in Pomeroy’s ratings.
Be selective picking upsets
Validation from picking upsets — defined here as a win by a team at least two seed lines below the losing team — is wonderful. I should know; I had fifth-seeded San Diego State in the title game last year, as you might have heard. Yet there probably aren’t as many of these upsets as you think, even in the early rounds. Since 1985, when the men’s field expanded to 64 teams, there have been, on average, 12 of these upsets per tournament.
Sometimes there are more — there were 14 last year, in a historically wacky tournament — and sometimes there are fewer — there were only four in 2007. As you would expect, the deeper you get into the tournament, the fewer such upsets occur. And obviously, if two lower-seeded teams meet in a later round, one will advance, even if it might not be a huge upset by seed.
So how do you decide which teams are capable of busting brackets? If you are comfortable with sports betting, check out the point spreads for each individual game and find lower-seeded teams that are either small underdogs or favored outright. Some of those this season include No. 10 Drake (which is favored over No. 7 Washington State), No. 11 New Mexico (which is favored over No. 6 Clemson) and No. 10 Nevada (which is favored over No. 7 Dayton), although there are also other appealing first-round upset candidates. And No. 11 Oregon is just a slight underdog against No. 6 South Carolina.
You could also check out the consensus rankings and make decisions accordingly. Historically speaking, the higher-rated team wins approximately 67 percent of the time, giving you a good indicator of a few potential upsets to target.
The size of your pool will tell you how much risk to take
If you are in a small pool — say, 25 people or less — you want to minimize your risk. That means not picking a large number of surprises, especially in the late rounds, and focusing on the favorites.
As the pool size gets bigger, it’s necessary to take more calculated risks to make your bracket unique. In other words, if you pick a favorite to win the title in a big pool — meaning you aren’t guaranteed much of anything even if your team wins — you need to be contrarian elsewhere, like picking a Cinderella team to reach the Final Four or taking more risks in the early rounds. If you decide to make riskier plays in the later rounds — the Elite Eight and beyond — you can play it safer leading up to the Sweet 16. Last year’s successful Perfect Bracket predicted that 12 of the top 16 seeds would advance to the Sweet 16; the riskier picks later on carried the bracket to success.
Don’t buy into the 12-seed mystique. Focus on the No. 11 seeds instead
You are going to hear a lot about how appealing the No. 12 seeds are, and how often they upset No. 5 seeds in the first round. That used to be the case, but lately there’s a better strategy. Since 2011, when the field expanded to 68 teams, No. 12 seeds have an 18-30 record against No. 5 seeds in the round of 64. The No. 11 seeds, by comparison, are 25-23 in their matchups against the No. 6 seeds — and often with a whole lot less hype. My top first-round upset picks this year include No. 11 New Mexico over No. 6 Clemson and No. 11 N.C. State over No. 6 Texas Tech.
Believe in at least one ‘First Four’ team
The First Four games open the tournament, pitting the last four automatic qualifiers against each other and the last four at-large teams against each other. The winners of the First Four games advance into the field of 64. They don’t all succeed there, but from 2011 to 2023, only once (in 2019) has an at-large First Four team failed to win a game in the 64-team bracket.
The most famous example is VCU steamrolling into the Final Four as a No. 11 seed in 2011. A decade later, UCLA went from the First Four to Final Four after beating No. 1 Michigan in the Elite Eight. La Salle (2013), Tennessee (2014) and Syracuse (2018) are other First Four teams with multiple wins in the 64-field tournament.
Last year got even crazier, when Fairleigh Dickinson became the first automatic qualifier to go from a First Four win to claiming a game in the round of 64. They did it in historic fashion, upsetting No. 1 Purdue to become just the second No. 16 seed to reach the round of 32.
And because these teams aren’t penciled into the 64-team grid on Sunday night, many bracket-pickers avoid them — which can give you an even bigger edge. This year, the “First Four” teams include four No. 10 seeds: Virginia plays Colorado State, and Boise State faces Colorado. Take a close look at the winners of those two games.
Favor teams that did well in conference tournaments
There was a time when you wanted your eventual title team to have won its just-concluded conference tournament, but that’s no longer necessary. From 1999 to 2010, eight out of 12 national champions previously won their conference tournaments. In the 12 tournaments since, just four conference champions won the national championship.
However, every national championship-winning team since 1985 — with the exception of UCLA in 1995 and Arizona in 1997, which didn’t have a conference tournament — has lasted at least to the semifinal round in its conference tournament. So plan on avoiding teams that made an early exit, at least for your national title pick. This year, such teams include No. 3 seed Creighton, No. 4 seed Duke, No. 4 seed Kansas and three teams from the SEC: No. 2 seed Tennessee, No. 3 seed Kentucky and No. 4 seed Alabama. Kansas at least had an excuse; Kevin McCullar Jr. and Hunter Dickinson missed the Jayhawks’ loss to Cincinnati because of injuries, and Parker Braun wasn’t 100 percent.
Focus on the statistics that matter
There are 68 teams in the tournament after thousands of games played this season, producing a variety of statistics indicating which teams will win or lose a particular matchup. Most of these are irrelevant. Instead, investigate the essentials of shooting, rebounding, creating turnovers and getting to the free throw line, also known as the four factors for offense and defense.
For upset candidates, offensive rebounding just might be the most important. Those rebounds provide teams with the extra possessions that are crucial to pulling off a March upset. Since 2011, in three-quarters of NCAA tournament upsets, the worse seed had the better offensive rebounding rate during the game in question. Some of the best offensive rebounding teams that are lower-seeded teams in this tournament include No. 16 Longwood (12th), No. 12 UAB (22nd) and No. 15 Saint Peter’s (23rd). Other strong offensive rebounding teams that could be under the radar include No. 9 seed Texas A&M, No. 5 seed Saint Mary’s and No. 7 seed Florida.
Don’t just guess at the tiebreaker total
The tiebreaker most often used — total points scored in the championship game — is often treated as an afterthought. It doesn’t have to be.
Since 1985, when the men’s tournament expanded to 64 teams, the national title game has averaged 145 total points when decided in regulation. The four overtime games in that span averaged 157 total points. The most total points scored in regulation was in 1990, when UNLV beat Duke, 103-73 (176). The fewest total points came in 2011, when Connecticut beat Butler, 53-41 (94).
How many points you choose should be influenced by which teams you have in the final, since pace of play and offensive efficiency help determine how many points a team might score. Here’s a quick list of some of the most frequent matchups in the Elite Eight and beyond and the average total points scored in those contests since 2011 — but look at the averages of your chosen teams. I will have a guide to picking the tiebreaker later this week.
Know how to spot a potential championship team
Success leaves clues, and we have a lot of data on how an eventual championship team usually performs leading up to the tournament. My colleague Matt Bonesteel outlines some of those clues in his annual best bets column — which included eventual national champion Connecticut as one of the five most likely winners last year — and there are some other guidelines you can follow. Since the field expanded to 68 teams in 2011, for example, every national champion except two — No. 7 seed Connecticut in 2014 and Connecticut again as a No. 4 seed last year — was a No. 1, 2 or 3 seed. Since 1985, when the field expanded to 64 teams, all but five of the 38 winners were a No. 1, 2 or 3 seed and 24 of the 38 (63 percent) were No. 1 seeds.
And all but four of the past 19 winners have had their individual Simple Rating System, a schedule-adjusted margin of victory rating that is expressed in points per game, rank in the top four nationally. Connecticut just missed last year, at No. 5. The top four schools in SRS this year are Houston, Arizona, Purdue and Connecticut.
You could also look at teams with similar profiles and see how far they advanced in the tournament, a technique that helped me identify San Diego State last year. For example, this year’s Auburn squad, a No. 4 seed, is similar to teams that have won a robust 2.3 games per tournament, on average (the same as No. 2 seeds from 2011 to 2023). This Auburn team also has similar performance metrics as runner-up Texas Tech in 2019 plus Florida in 2017 and Houston in 2022, two Elite Eight teams.
