NCAA: Advice from a gambler on getting your bracket right

(Bloomberg View) -- The National Collegiate Athletic Association basketball tournament begins Thursday,  so it's time for bracket pools.  Yet most people approach their bracket picks in a flawed way, and these errors are much like those amateurs and professionals make in investment decisions.

Only upset picks matter.  If you make too few of these, you're unlikely to win the pool even if you all your choices are correct. If you make too many  picks, you're unlikely even to have more upset wins than losses. The key is picking the right number.

Many bracket pool participants try to guess which teams are likely to win. This method doesn't help much.  It's possible that by searching through all the games an expert handicapper could increase the chances of winning by 10 percent or 20 percent, not much more.

Another popular approach is trying to pick outcomes that others don't. Here too, there's not much advantage even if you're successful.  If you know your fellow participants well and can find some good picks they are unlikely to make, you might increase your chance of winning by 50 percent, but it's hard to imagine a lot more.

Getting the right number of upset picks, on the other hand, makes a huge difference. Most people are essentially eliminated from winning their pools before the first game is played because they have too few or too many of these selections. But guessing the right number of upset picks can give you about one chance in the square root of the number of participants of winning. So a 20 percent chance of winning a pool with 25 participants, a 10 percent chance if there are 100 people and a 1 percent chance if there are 10,000 entries.

How does this apply to finance? The things most people focus on -- which teams win games, which picks are made by participants in the pool -- are outside their control. Better to concentrate on things you do control, such as how many upset picks you make (of course, you have no control over how many will pan out, but you can choose how many tries to make). Similarly, a lot of investment commentary is devoted to trying to guess what the stock market or the Federal Reserve might do. The main choice that determines financial security for individuals is the fraction of income they save, and that is entirely under individual control. Pension funds will succeed or fail mostly as a result of funding status, which at least theoretically is under political control.

The emphasis on estimating the win probabilities in the NCAA tournament is similar to the focus on estimating the expected return of assets. These are both hard problems, partly because they are complex, but mainly because the market consensus is difficult to improve upon. Maybe individual and institutional investors can do a little better by guessing future expected returns, but there is far more benefit to forecasting future risk. In bracket pools that means determining the number of upset picks. In investing it means picking the right target volatility. In both cases, this is an easier problem, with greater payoff.

Worrying what other people will pick in bracket pools is like investors searching for overlooked assets, or thinking about the effects of people crowding into and out of trades. This is another hard problem, because people are unpredictable. It's smarter to insulate yourself from the effects of fads and fallacies, which means picking your investments based on solid principles -- whether those are backed by theory and empirical evidence or by sound fundamental research -- rather than trying to guess what other people think.

You can live your life making conventional choices, doing what everyone expects. That will get you an outcome that you cannot easily predict, as it depends on your health, romantic luck, the economy, natural disasters, political events and other stuff. You can spend a lot of time worrying about how things will turn out if you do the usual things; but that's just the hand you were dealt.

It's the unusual choices, the upset picks, that play your hand. These are the things you control, the things that will determine whether you do better or worse than the easy path laid out for you.

Bracket pools are mostly for fun, though as a lifelong gambler I cannot help treating any contest with money payoffs seriously. But financial decisions can make a tremendous difference to individual happiness and institutional survival. If you must worry about things you cannot control, focus on stuff that's hard to predict and doesn't make much difference rather than stuff that's easy to predict and matters a lot; and spend your energy on the favorite picks that most people make rather than the upsets that determine outcomes -- but do it in your bracket pools, not your financial life.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Aaron Brown is a former Managing Director and Head of Financial Market Research at AQR Capital Management. He is the author of "The Poker Face of Wall Street."

  1. Not counting the “first four” play-in games starting Tuesday.

  2. The popular pool contest requires participants to pick winners in each of the games that determine the national champion.

  3. For every game there is a favorite team, and favorites win about two-thirds of the time. Let's say favorites win F of the games (about in a typical year) and in your bracket picks you make U upset picks and 63-U favorite picks. If you win W of your upset picks it means you win F+2W-U total picks. F is the same for everyone in the pool, so the winner will be the participant with the highest value for 2W-U, which can also be described as upset pick wins minus upset pick losses. This simplifies real pools in which there is more than one possible upset winner in each game after the first round, and scoring is often more complex than just adding up the number of correct picks. But the general point remains true for more detailed mathematical analysis.

  4. Remember, each one has less than a percent chance of winning.

  5. Suppose you knew for sure which team would win a game. This gives you the same advantage as someone who makes two identical pool entries except picking different teams for the one game. But those two pool entries would have less than twice the chance of winning as the entry with the better chance, both because one entry has a better chance than the other, and because the entries have a good chance of either both winning or neither. Therefore, percent knowledge of one game increases the value of your selections by less than percent. But you're not likely to have percent knowledge of any game, even knowing one team has a percent better or worse chance than consensus is pretty rare.

  6. Using a similar argument to the one above, if there is an upset pick that percent of participants should make playing optimally, and you know for sure that no one else will pick it, making that pick increases your chance of winning by less than percent. Given that it's hard to know for sure what other people will do, you need to know a lot about your fellow pool members to get much advantage from guessing what they'll do.

  7. Other important choices are amount of financial risk to take and degree of diversification.

To contact the author of this story: Aaron Brown at

For more columns from Bloomberg View, visit

©2018 Bloomberg L.P.