March Madness has arrived, which means you’re going to hear a lot of bracket-picking advice from all corners of the internet and the airwaves. But there are no golden rules for picking an NCAA bracket that gives you the best chance to win. Worse yet, a lot of conventional wisdom can get you into trouble.
Based on more than 15 years of data-driven research into bracket strategy, here are three common pitfalls you need to avoid to give yourself an edge in your NCAA bracket pool.
Not Picking A Team Because Of Past Tournament Performance
Recency bias is real, and it has significant implications for bracket-picking strategy. People often are scared to choose a team that has choked in recent NCAA tournaments, which can sometimes result in a top contender being an unpopular pick for reasons completely unrelated to this year’s performance.
You will find no stronger illustration of this principle than Virginia winning the national title in 2019. Just a year earlier, Virginia was the first No. 1 seed in tournament history to lose to a No. 16 seed in the first round. The Cavaliers also had been a high seed in four of the previous five NCAA tournaments but had come up short of the Final Four each time.
By the start of the 2019 tournament, Virginia again was one of the leading title contenders and a No. 1 seed. Yet the Cavaliers remained a great value pick to make the Final Four and win the title. That means the percentage of the public picking Virginia to advance to the final rounds was lower than its actual odds to do so, according to objective computer predictions and the betting markets.
Our research shows that such recency bias extends far beyond this example. In fact, over the last decade, the previous year’s tournament underperformers were consistently good value bracket selections compared with teams that had overperformed expectations the previous year.
For example, prior-year underperformers — defined as teams that won at least one fewer game than expected in the previous NCAA tournament — were picked to reach the Sweet 16 by 33% of the public. But a significantly higher 41% of those teams made it.
On the other hand, the public selected prior-year overperformers — teams that had at least one more win than expected in the previous NCAA tournament — to reach the Sweet 16 36% of the time. Yet only 25% made it.
Finally, here’s how the last 10 national title winners did in the previous year’s NCAA tournament:
— Two didn’t even make the tournament.
— Four failed to reach the Sweet 16 as top seeds.
— Another lost in the Sweet 16 as one of the tournament favorites.
— Only three made deep runs.
Overcoming this bias is psychologically tough. If you pick a team that burned you last year and it ends up burning you again, it’s the absolute worst feeling. But if you want to maximize your long-term edge in bracket pools, you need to avoid this trap.
Crafting A Bracket To Match Historical Seed-Based Trends
Every NCAA tournament is different. Some years the top few teams are clearly a cut above the rest of the field. In other years more parity exists across the top two or three seed lines. And occasionally the selection committee just completely whiffs, as when national finalist Kentucky got a No. 8 seed in 2014.
The point is that from tournament to tournament, team quality at a particular seed line can show stark differences, so bracket advice that considers only seed numbers is suboptimal. Yet we’ve come across some truly mind-boggling bracket advice along these lines. In 2019, for instance, one pundit recommended picking a Final Four in which the seed numbers added up to 11, because that was the average for Final Fours.
Funny enough, the seed numbers of the Final Four teams in 2019 actually did add up to 11: No. 1 Virginia, No. 2 Michigan State, No. 3 Texas Tech and No. 5 Auburn. However, what do you think the odds were that someone following that “seed sum of 11” advice would have nailed that specific combination of teams? Close to zero.
Virginia was an unpopular top seed, Michigan State was in the same region as tournament favorite Duke, and Auburn and Texas Tech were very unpopular picks. Consequently, a likely outcome of the “seed sum of 11” would have been getting no Final Four selections correct.
Picking more conservatively — for instance, a Final Four of all Nos. 1 and 2 seeds — would have likely resulted in better scores on average.
By trying to match seed-based historical performance trends, you also end up picking too many upsets, which is the silent killer of many brackets. Upsets will happen every year, but that doesn’t mean the best strategy is to try to predict them all. Even winning brackets get plenty of picks wrong.
The “12-seed over 5-seed” advice you hear every March is a good example. On average, at least one 12th seed per year has pulled an upset in recent tournaments. So you should pick at least one 12th-seed upset, right?
No, probably not. Think about it this way. Let’s say you pick all the No. 5 seeds, while four of your bracket-pool opponents each select a different 12th-seed upset. The first round tips off, and yet again, three No. 5 seeds win and one No. 12 seed pulls an upset.
In that case, your total score from those four games will be better than three of those four opponents. Repeat that dynamic over several tournament rounds and lots of other seed lines, and the cumulative edge of not chasing every upset really adds up.
There is a saying, often credited to Voltaire, that “the perfect is the enemy of the good.” When it comes to bracket strategy, a lot of players strive for perfection and make too many risky picks in doing so. The goal is not to choose a perfect bracket. The goal is to get a higher score than your opponents. Correctly picking all the long shots that actually win is incredibly difficult, and attempting to do so is usually a great way to fall behind in the standings.
Focusing Too Much On Value And Not Enough On Overall Pick Risk
Finally, even educated bracket pickers can fall too much in love with the concept of finding value picks and diversifying their bracket from the public. In the last several years, awareness of this strategy has mushroomed, to the point that some consider it gospel to “pick an underrated team as your NCAA champion.”
At a high level, it’s generally good advice, but the logic has limits. For example, if a team has a 10% chance to win the national title but is being picked to do so by only 6% of the public, that is technically a value pick.
For the sake of example, let’s say every other team with a better chance of winning the tournament, including one big favorite with 40% odds to win it all, is being overvalued by the public. So this team with a 10% chance is the strongest undervalued-champion pick in the NCAA field. Should you pick that team?
It depends. In a larger bracket pool, that team likely would be a solid choice. But in a smaller pool, not picking the odds-on favorite to win — and giving up a 40% chance to earn 32 points — could actually lower your odds to win the pool, even if the public is overvaluing the favorite. It’s simply too big a risk to take in a small pool, and you could always find more suitable value pick opportunities in earlier rounds.
In short, you need to remember that every time you don’t choose the odds-on favorite to make a particular round, you’re reducing your expected total score. So in taking added risk to make a value pick, you need to pick your spots.
Figuring out that perfect balance between risk and value isn’t easy. The optimal picks for any pool depend heavily on factors like the number of entries, the scoring system and how you expect your opponents to fill out their brackets. It requires a lot of data and math to get the best answer, which is why we built technology to do all the necessary number-crunching.
Still, just keeping these three pitfalls in mind will improve your odds to win a pool. Don’t let the ghosts of past tournaments cloud your judgment, ignore seed-based pick advice and spend more time thinking about the overall level of pick risk appropriate for your pool than agonizing over which first-round upsets to pick.
TeamRankings.com has provided expert bracket-picking advice since 2004.
Since 2017, TeamRankings subscribers have reported over $1.7 million in bracket-pool prize winnings using the NCAA Bracket Picks.
Want more? Get customized picks for your pools and in-depth analysis of the 2021 NCAA bracket with TeamRankings.com’s NCAA Bracket Picks.
VSiN readers interested in subscribing can also get an exclusive discount.