You want to win your March Madness pool.

The money is nice but the bragging rights are even better, especially because of Josh. He played college basketball and loves to wax on about the 3.7 points per game he scored his senior season at Harrison University.

Josh has spent the last two months watching college basketball to fill out his bracket. How can you compete?

By being smart.

In 2015, I wrote a slim book called How to Win Your NCAA Tournament Pool. It gave numerical proof that contrarian strategies increase your odds of winning certain types of pools. Above a certain threshold of people in your pool, you have to fade what others are doing.

Over the past decade, I’ve refined my thinking on how to win your pool as my ideas have appeared in major media outlets such as Yahoo. In these top 5 tips, I summarize key points from the original book as well as my latest advice for beating Josh.

Let’s get into it.

1. The one thing you must get right

Predictive analytics is crucial in helping you pick winners in your bracket. We’ll soon get into the specifics of the accuracy of the predictions here at The Power Rank. However, there is more to winning your pool than analytics.

In fact, if you get this one thing wrong, analytics will not help. Unless you’re Biff from Back to the Future Part II, you might as well light your entry fee on fire.

To understand the one thing you must get right to win your March Madness pool, consider an analogy.

Steph Curry walks into a gym. There are a bunch of college players looking to get into the NBA. Steph is feeling generous, so he sets up a contest. Beat Steph in a three point shooting contest, get an NBA rookie contract.

As the greatest shooter to ever grace a basketball court, Steph goes out and sinks 17 of 25 three point shots, an amazing mark. Can any college kid beat him?

You might think Steph will certainly win the contest. However, it depends on the numbers of college kids in the shooting contest.

For every college player that steps up, let’s assume a 95% chance that he makes fewer shots than Steph. If two players enter the contest, the chance that Steph beats both is 0.95 times 0.95. He still has a healthy 90% chance to win the contest.

However, Steph’s odds to win decrease with every additional contestant, as his win probability gets another multiplicative factor of 0.95. If 13 college players participate, the odds are about 50-50 that Steph wins. This goes down to 7.6% with 50 college players.

If you use my analytics in your March Madness pool, you’re like Steph in this shooting contest. You have great odds to beat any one person.

However, your odds go down with more people in your pool. The bigger the pool, the more likely someone gets lucky and beats you.

Someone like Grandma. She doesn’t know anything about college basketball. However, she’s Catholic, and that matters in 2018.

Grandma has Villanova, the oldest Catholic university in Pennsylvania, as champion. In addition, she’s also friends with Sister Jean and picks Loyola of Chicago to make the Final Four. 

When Loyola of Chicago made the Final Four, Grandma got points that no one else earned from the South Region. When Villanova beats Michigan in the finals, Grandma won her pool. It’s as if Shaq beats Steph Curry in a three point shooting contest.

Here’s the take home message: Do not get in a large pool. For winner take all contests, I recommend a pool with less than 100 people. Maximize your odds of winning by getting in a pool of 10 or less.

2. The simple, smart way to fill out your bracket

Now that we’re in the right pool, let’s look at predictive analytics.

There are many strong resources out there, from the veterans like Ken Pomeroy to upstarts like Evan Miyakawa. Let me use a huge data set to make the case for my own.

My college basketball team rankings at The Power Rank take margin of victory in games and adjust for schedule with my mathematical algorithms. Let’s look at these team rankings before the start of every tournament since 2002. The higher ranked team has won 70.9% of tournament games (1072-441).

For comparison, consider the closing market, the gold standard for prediction. I use data from the provider KO STATS. The market favorite wins tournament games at a 71.4% rate (counting 19 market pushes as half a win).

Coming close to beating the market is impressive as my numbers do not benefit from any tournament games in making a prediction. The market can capitalize on the latest tournament games as well as the collective insights of sharp sports bettors.

Why does the algorithm do so well at predicting the results of tournament games? It skips wins and losses and instead looks at margin of victory in making an accurate prediction of team strength. 

Every day, my Apple Macbook Air takes all the game results for the season. The algorithm cranks through every single margin of victory and then makes mathematical schedule adjustments. This requires solving 365 equations for 365 variables, one for each college basketball team.

Let’s also look at how the team rankings do in predicting the winner of the NCAA tournament. The choice of champion is the most important choice in your bracket, worth 32 points in most pools.

Over the past 23 tournaments, the winner has ranked first or second in 13 of those years, more than half. There were another 5 tournaments won by a team ranked in the top 5. This means the champion was ranked in the top 5 of my pre-tournament rankings in 78% of tournaments.

Last year in the 2025 tournament, Florida was a healthy number one in my college basketball rankings over Duke. Todd Golden’s team needed some heroics from Walter Clayton Jr. against Houston in the final but won a national championship.

It won’t work out this nicely every year. However, my March Madness cheat sheet is based on these team rankings at The Power Rank and makes it drop dead easy to fill out your bracket.

To get this delivered straight to your inbox the day before the tournament starts, enter your best email and click on “Sign up now!”

 

 

 

3. The seemingly stupid but actually powerful predictor of NCAA tournament games

Data is king. You’ve already seen the predictive power of my college basketball team rankings at The Power Rank.

But what if I recommended an alternative to data that also makes remarkable tournament predictions? This unorthodox method also works in the NFL postseason. Yes, it ignores all game data from the current season.

But when I tell you that this predictor is the preseason AP poll, you might think that’s the stupidest thing you’ve ever heard. How can a bunch of sports writers make accurate tournament predictions before the season starts?

Before we get into the reason, let’s look at some results. From the 2002 through 2025 tournaments, the higher ranked team in the preseason AP poll won 71.6% of games (934-370).

In this calculation, teams outside of the top 25 get ranked based on points. A team gets 1 point for getting ranked 25th in a ballot, 2 for 24th and so on. I exclude teams with less than five total points.

Ranked teams are predicted to beat unranked teams, but this still leaves a fraction of games with no prediction between two unranked teams. For comparison, let’s look at the accuracy of the closing market in the same games that the AP poll makes a prediction.

In these games, the closing market predicts the winner in 73.0% of those games (15 market pushes counted as half a win). I find it remarkable that the preseason AP poll gets within one and a half percent of the gold standard of prediction without any data from the regular season much less the NCAA tournament.

The preseason AP poll works because of the wisdom of crowds. No one sports writer has the perfect ballot. However, the aggregate of many ballots leads to a powerful predictor of team strength that persists into the NCAA tournament.

Every year, I create a cheat sheet for members of The Power Rank to fill out a bracket. Unlike the cheat sheet in the newsletter, these prediction use my member college basketball numbers that use market as well as game data. In addition, I mention the preseason AP poll if it differs from my model. The data suggests to take it seriously.

Let’s not stop there, as the preseason AP poll has an uncanny ability to highlight overrated teams. 

From 1985 to 2025, 40 teams started the season outside the top 25 in the preseason AP poll but then earned a 1 or 2 seed in the NCAA tournament. These teams exceeded preseason expectations by a large amount.

However, none of these 40 teams has made the Final Four. Zero. Zilch.

Let’s assume there is about a 20% chance that a typical 2 seed makes the Final Four, which I estimated from my projections over the years. It is also the rate at which 2 seeds have made the Final Four since the tournament expanded to 64 teams in 1985. There is a 1 in about 7500 chance that none of these 40 teams makes the Final Four at random.

The preseason AP poll has an uncanny ability to highlight overrated teams. 

4. How to predict upsets

Upsets make March Madness special. 

In 2018, Virginia was the top overall seed heading into the tournament. They got shocked by UMBC, as a 16 seed beat a 1 seed for the first time in men’s NCAA tournament history.

The only thing better than watching these upsets is picking them in your bracket. How can we use analytics to do that?

Maybe others have some insight. For the past two seasons, The Athletic has written about Bracket Breakers, a math model that predicts March Madness upsets.

Let’s focus on one result: teams that pull off the upset tend to shoot a lot of threes. This makes sense, as shooting more threes should increase the variance in points. An increased variance favors the underdog.

I looked into this result but not only on tournament games. I considered all college basketball games in which a team closed as an underdog of six or more points but won outright, a much larger sample size.

During the four seasons prior to the 2025-26 season, college basketball teams took about 38% of their field goal attempts from three. Underdogs of six or more points that won had a three point rate 1.0% lower. Pulling off the upset is not about shooting more threes.

In addition, I’ve found no statistical relationship between a team’s three point rate and variance in points per possession. The idea that shooting more threes increases variance is not supported by the data.

Then what leads to upsets? Making three pointers.

During the four seasons prior to the 2025-26 season, underdogs of six or more points that won made 5.3% more of their threes than their season average. The corresponding favorites that lost made 5.2% less from behind the arc.

This difference in three point shooting for both teams resulted in about a seven point swing. This was by far the largest effect that I found.

For example, there is also an effect from two point field goals. Like from behind the arc, underdogs make a higher rate of two pointers while favorites make a lower rate. This resulted in a two point benefit for the underdog.

Fantastic. Three point shooting leads to upsets. All we need to do is predict three point field goal percentage.

However, this is difficult. In a landmark study called The 3-point line is a lottery, Ken Pomeroy found no correlation in a team’s three point percentage from early to late season. This was true for both offense and defense.

The results on offense should be particularly surprising. Shooting is a skill, right? How can it not be predictive?

That’s a nuanced question for another day. For March Madness, let’s conclude that the randomness of three point shooting makes it difficult to predict upsets. This is another reason to go with the analytics. But you might not.

5. How your brain prevents an optimal bracket

To give yourself an edge in your pool, you know to use analytics. As explained earlier, my pre-tournament numbers have been almost as good as the markets for more than two decades.

In addition, filling out a pool with the higher ranked team in my college basketball rankings is easy. You’ll get my cheat sheet the Wednesday before the start of the tournament, and you just fill in the winners in each region.

However, you might not take this advice. You would have two complaints:

• You won’t get every game correct.
• Your bracket is boring.

Both of these things are true, and it causes you to consider that 12 over 5 upset even though it goes against my cheat sheet. It’s March, the one time of year to find that matchup that might lead to an upset.

There is some interesting brain science behind why you might not follow the cheat sheet. In the book The Ravenous Brain, Daniel Bor describes a light detection experiment. 

Participants are asked whether the left or right light will flash next. The left light flashes with 80% probability at random, and the right flashes the other times.

With a few lines of Python code, I generated a sample sequence with 80% chance for left:

L – L – L – L – L – R – L – L – L – L

In the experiment, it is optimal to guess left every time. Rats figure this out.

However, humans do not. We try to get every flash correct and guess left 80% of the time. This decreases the accuracy of the predictions.

Think about the sample sequence above with one right flash. If you guess left every time, you will get 9 of 10 light flashes correct. If you pick two rights at random, you will at most get 9 of 10 correct.

I find this experiment horrifying. Suppose you had a system that picked 56% of winners against the spread. With this win rate and proper bankroll management, you will grow your wealth at an impressive rate.

However, the light detection experiment shows you can second guess this winning system. Let’s assume you bet your system 56% of the time at random, choosing to go against it the other times. Your win rate drops to 50.7%, an unprofitable winning percentage.

You might also second guess my cheat sheet and consider a 12 over 5 seed upset in the Round of 64 that goes against my numbers. You’re almost a communist if you do not pick an upset like this. However, the light detection experiment says that you should not do this.

Be like a rat. Use analytics to fill out your bracket.

This tool makes it drop dead easy to fill out your bracket

If you sign up for my free sports betting newsletter, you’ll get my March Madness cheat sheet in which I lay out every predicted winner in an easy to use format. It’s based on my college basketball team rankings in which the higher ranked team has won 70.9% of games the last 23 tournaments.

And remember those contrarian strategies? You can also download a free pdf of my book How to Win Your NCAA Tournament Pool in which I discuss those ideas in Chapter 3.

To get these March Madness goodies, enter your best email address and click on “Sign up now!”