Bracket-picking strategy: how to win a pool
How to fill out an NCAA tournament bracket to win a pool rather than to be right: scoring rules, pool size, upset selection, champion leverage, tiebreakers and the mental traps that sink most entries.
There are two ways to fill out a bracket. One is to try to predict the tournament, picking the most likely winner of every game and hoping to be close to perfect. The other is to try to win your pool, which is a different goal and often calls for a different bracket. The first approach feels natural, and it is what almost everyone does. The second is what people who win office pools consistently do, whether they know it or not.
The difference comes from the structure of the competition. In a pool, you are not scored against the tournament. You are scored against the other entries. A bracket that is very accurate but looks exactly like everyone else's will tie with the crowd, and a tie in a large pool means a very small prize or none. A bracket that takes a few well-chosen risks can finish first on the days when those risks pay off. Good bracket strategy is therefore partly about forecasting and partly about game theory, the study of how your choices interact with other people's.
This article lays out a practical approach. It covers what to find out about your pool before you pick, how scoring changes the value of each game, how pool size should change your picks, how to choose a champion, where to place your upsets, how to handle tiebreakers, and which mental traps to avoid. It is educational and deliberately general: it does not pick teams, because the right picks depend on each year's field. If your pool involves money, remember that the rules about wagering differ by place, that participation is for adults only, and that an office pool should be fun first.
Start with the rules of your pool
Before you pick a single team, read the scoring system and count the entries. These two facts drive nearly everything else.
The most common scoring system doubles the value of a correct pick each round. A typical version awards 1 point per correct pick in the first round, 2 in the second, 4 in the third, 8 in the fourth, 16 in the national semifinals and 32 for the champion. Some pools use 10, 20, 40, 80, 160 and 320, which is the same thing multiplied by ten. Under this system, each round is worth exactly the same total: 32 first-round games at 1 point is 32, 16 games at 2 is 32, and so on. A perfect bracket scores 192 in the 1-2-4-8-16-32 version.
That equal weighting by round has a consequence that surprises many people: the late rounds are worth as much in total as the early rounds, but they are determined by far fewer picks. A single decision, your national champion, is worth one-sixth of the entire pool. The four Final Four picks together are worth another sixth. If you get the early rounds right and miss the late ones, you will usually lose to someone who did the opposite.
Other systems change the calculation.
- Upset bonuses. Some pools award extra points equal to, or based on, the seed of the winner. These make longshot picks more valuable and reward risk.
- Seed-weighted scoring. A pool that multiplies each pick's value by the winner's seed makes the late rounds with lower seeds more lucrative.
- Straight points per game. A system that gives every game the same value makes the first round far more important than in the doubling system, because there are so many games.
- Confidence pools. These ask you to rank your picks and assign points accordingly. They reward accurate ordering of your own certainty.
Check also how the tiebreaker works, whether you may enter more than once, and whether there are prizes for second place or for a single round. All of these change the best strategy.
Pool size changes everything
The most important strategic idea in this article is simple: the larger the pool, the more you must differ from the crowd.
In a small pool of eight or ten people, most of whom are casual fans, you can win by being sensible. If you pick the favorites in most places and make a few careful upsets, you will usually finish near the top, and a few of the other entries will contain obvious mistakes such as picking their own school to win the title. Your goal is to maximize your expected score.
In a large pool of hundreds or thousands of entries, the prize goes to the very top of the standings. The best entry in a thousand is not the one with the highest expected score. It is the one that scores highest in the particular tournament that happens. To finish first, your bracket has to be right in places where most other brackets are wrong. If you pick the same champion as most of the pool and that team wins, you share the glory with the crowd. If you pick a less popular champion who has a real chance and that team wins, you are alone at the top.
This is called leverage. A pick has leverage when it is different from what most entrants choose and has a reasonable chance of being right. The best picks are not the ones with the highest chance of coming true. They are the ones where the chance of coming true is higher than the share of entries that chose them. Imagine two potential champions. The first is likely to win with a probability of 20 percent, but 40 percent of the pool has picked it. The second is likely to win at 12 percent, but only 4 percent of the pool picked it. If the second team wins, you are far better placed. You are giving up some chance of being right for a much bigger payoff when you are.
You do not need to know the exact public pick percentages to use this idea. A rough rule is that the public loves top seeds, famous programs, local teams and recent winners. It tends to underrate good teams from lesser-known programs, and it overrates teams that have been in the news. Knowing the popular pick helps you find where to differ.
For the casual-fan version of these ideas, our companion piece on college bracket pools covers the basics, and our notes on betting and the NCAA tournament cover how the pool dynamics differ from other contests.
Choosing a champion
The champion pick is the most valuable single decision, and it deserves the most care. History helps narrow the choice.
Since the tournament began seeding teams, almost all of the champions have come from the top few seed lines. Most have been 1 seeds, and nearly all the rest have been 2 or 3 seeds. A handful of exceptions have come from lower lines: NC State as a 6 seed in 1983, Villanova as an 8 in 1985, Kansas as a 6 in 1988, Arizona as a 4 in 1997, UConn as a 7 in 2014 and UConn as a 4 in 2023. The exceptions are real, and each one rewarded the few people who picked it, but they are rare enough that a pool entry that picks a champion outside the top three lines is taking a very large risk. The history is laid out in the history of the one-seed.
A sensible process is as follows.
- Make a short list. Start with the teams in the top two or three seed lines and cut those with obvious weaknesses, such as a bad defense or a limited rotation.
- Check the ratings. Efficiency ratings, which we describe in college efficiency metrics, are the best guide to team quality. A team whose rating is far above its seed is a candidate for leverage, and a team whose seed is well above its rating is a candidate to avoid.
- Look at the path. Note which regional opponents the team would face. A top team in a region with another very strong team is more likely to be eliminated before the Final Four.
- Consider the pool. In a small pool, choose the team that is most likely to win. In a large pool, choose from the short list the team that is least popular among your fellow entrants.
- Check the injury news. If a key player has been hurt, the rating may not reflect it. Make sure you check the latest reports just before you submit.
Remember also what the Final Four looks like. It is rare for all four 1 seeds to reach it, and that has happened only twice in the history of the 64-team and larger fields, in 2008 and 2025. On average, between one and two 1 seeds reach the Final Four in a typical year. So a bracket that advances all four 1 seeds to the national semifinals is making a very unusual claim. If you pick two 1 seeds and place the others from the 2, 3 and 4 lines, you are much closer to the historical shape.
Where to put your upsets
Upsets are the spice of a bracket and the main way most people lose their pools. The key is to choose them carefully and to place them where they hurt least.
Pick a few, not many
An upset pick is a bet that a team with a lower probability will win, and the more you make, the more likely you are to be wrong somewhere. The first round has 32 games, and a typical year contains several upsets by seed, many of them in the 8-9 games, which are close to coin flips. But any particular upset is unlikely. Choosing a handful of them in the places where the odds are best, and no more, balances the risk. We explain where those odds sit in why upsets happen.
Favor the right matchups
The most common first-round upsets are the 12 over 5 and the 11 over 6, and the 10 over 7 and 9 over 8 are nearer to even. The 13 over 4 and 14 over 3 happen but are much rarer. The 15 over 2 and 16 over 1 are so uncommon that picking them is almost never worth it. For 12-5 specifically, the pattern is that 12 seeds win a bit more than a third of the time, so picking one or two of the four is sensible. We cover that in the 12-5 upset: history, odds and the pattern.
Place them away from your deep picks
This is the most underused idea in bracket strategy. When you pick an upset, you do not just risk the points for that game. You also remove the favorite from your bracket. If you had that favorite going deep, you lose all the later points as well, and if the upset does not happen, you lose the points for every round that favorite would have played.
The cost of a wrong upset pick is therefore much larger when the favorite you are eliminating is one of your Final Four or champion picks. The practical rule is to put your upsets on the games where the favorite is not part of your deep run. If your champion is a 1 seed in a given region, do not pick an upset over a team that the 1 seed would face. Make your upsets in the other parts of the bracket, where you have nothing riding on the favorite.
The reverse is also true. An upset in the second round that eliminates a 1 seed from someone else's champion pick will cost them dearly, and your entry gains a relative advantage. In a large pool, that is exactly where your leverage lies.
Keep your second-round picks consistent
A related point is to think of your bracket as a set of linked decisions rather than a list of independent ones. If you pick a 12 seed to win its first game, you also need to decide whether it can win its second, because the 12 would then face the winner of the 4-13 game. Most picks for a Cinderella should stop after a game or two, because a team that is lucky once is rarely lucky twice. See Cinderellas: the greatest runs by double-digit seeds for how rare long runs are.
The tiebreaker
Most pools break ties by asking for the combined score of the championship game. It looks like an afterthought, but in a large pool it can decide the winner, since many entries with the same champion and similar picks finish with the same number of points.
The best approach is to pick a number that is plausible. Many entrants guess a round figure near the middle of the range, so a number a little away from the obvious one gives you a chance to be the closest when the crowd clusters. In recent years, most combined totals in the title game have landed between about 130 and 160, with the lower and higher ends rarer. A guess in the 130s or low 140s is reasonable. Check the exact rule in your pool: some tiebreakers are the closest total, and others punish going over.
Mental traps that sink brackets
Most entries do not lose because of a lack of information. They lose because of predictable mistakes. Here are the main ones.
- Loyalty. People pick their own school or their favorite team too high. If you love a team, you can still pick it, but ask what you would pick if you had no feelings about it.
- Recency. A team that played well in its conference tournament or got a win over a famous opponent last week looks better than its season suggests. Weigh the whole body of evidence.
- Narrative. The team with the best story, the best coach interview or the most fans is not always the best team. Check the ratings.
- Chasing upsets. A bracket full of upsets feels clever, but it is a very high-variance bet that nearly always fails. The pool does not reward you for being bold. It rewards you for being right in the right places.
- Being too safe. The opposite error is just as common. A bracket of all favorites is almost certain to lose in a big pool.
- Overreacting to a single result. A team that lost a game in the last week of the season is not necessarily worse than it was. Games are noisy, as we have seen.
- Peeking and copying. Looking at what others have picked can be useful for leverage, but copying a popular bracket removes your chance to differ.
The women's bracket
The same principles apply to the women's tournament, with one adjustment. Historically, the women's event has produced fewer upsets at the top of the bracket than the men's. A 16 seed beating a 1 seed has been far rarer there: Harvard's win over Stanford in 1998 stood as the only one for decades, and the top seeds have dominated the Final Four in most years. The consequence for pools is that chalk, picking favorites, is a sounder base for the early rounds, and the leverage comes in the champion and Final Four selections, where a few strong programs compete. We discuss how the women's event works in the women's NCAA tournament: format, regions and the growth of the event. You can see both brackets on the site in the men's bracket and women's bracket pages.
A simple recipe
Here is the whole approach in a few steps that you can apply to any year's bracket.
- Read the scoring and count the entries.
- Rank the contenders for the title with efficiency ratings, and note the injuries.
- Choose your champion from the top three seed lines, favoring a less popular option if the pool is large.
- Choose a Final Four with no more than two 1 seeds unless you have a strong reason.
- Pick a few first-round upsets, mostly 12 over 5, 11 over 6 and 10 over 7, in parts of the bracket where you do not have a deep run riding on the favorite.
- Choose a tiebreaker number in a sensible range.
- Check the latest news, then submit and stop tinkering.
What to take away
Winning a pool is not the same as predicting the tournament. The scoring rewards the late rounds, the size of the pool determines how much you must differ from the crowd, and your upsets cost the most when they remove a favorite you had going deep. The best brackets combine good forecasting, from ratings and history, with an awareness of what everyone else is likely to pick.
Above all, keep the contest in proportion. Even the best-reasoned bracket will be wrecked by something nobody saw coming, because that is the nature of the event, and no strategy can remove the randomness that makes it worth watching. The point of the strategy is to give yourself the best chance, not to guarantee a result.