14 things you need to know to win your college basketball pool

The PickHoops staff has seen many thousands of pools, half a million brackets, and has decades of combined pool experience. Here's some of what we've learned.

1. Establish expectations: you're going to need luck

Don't kid yourself. Predicting the outcome of a basketball game cannot be done with certainty. If you're in a pool with 50 people, you have to be in the top 2% to win your pool. You may consider yourself very good or even clairvoyant because you picked a few upsets last year, but luck is rarely sustained over time.

Bottom line, it is going to take a some luck to win any bracket contest. But there are some things you can do to take best advantage of your luck.

2. Know your goal: beat the no-upset bracket

We've established that you're going to need some luck, and that means doing better than a bracket that picks no upsets. Any fool can pick a no-upset bracket, and many do, so consider that someone in your pool will.

The no-upset bracket is going to beat roughly 85% of human picked brackets in a given year. You're really competing against the other "lucky" 15%.

3. Know your pool size: how many is 15%?

If your pool is small (10-20 or so), the lucky 15% is 2 or 3 players. Play it to beat only them, so that means a bare minimum of solid upsets.

If your pool is medium sized (50-100 or so), you're competing against a dozen people. You need to pick a key upset or two or three to win, but don't pick too many.

If you're in a massive pool (many hundreds and up), you need a miracle to win. And you have to pick for your miracle to happen. Your semifinal picks have to be exact or nearly so. Pick realistically but leave no upset unturned. Pick for something unexpected to happen, because something always does.

4. Know your scoring system

The simplest kind of scoring has a flat number of points per round, offering no bonus for correctly picking an upset. In this scoring system, you should not overdo it in picking upsets.

There are many different types of bonus scoring systems. Those that offer bonuses for upsets allow you to be much more generous in your selection of upsets. More on how to take advantage of this later.

5. Make sure your chosen winner is likely to be a contestant.

This is especially true in rounds 2 through 4. You might look at a 4th seed and think they have a good shot at beating the 1st seed in round 3. But don't forget that the 1st seed has a 87% chance of playing in round 3, while the 4th seed, on average has a 43% of playing. You don't (generally) get points for predicting who will lose. You only get points for selecting who will win. You can't afford to have many games where your winning selection doesn't even play. Now if the 1st seed falls to the 16th seed, you don't have to worry as much because most everyone else in the pool will get that pick wrong too.

6. Lean away from local favorites.

Unless your office pool has entrants distributed throughout the country, the contestants will generally tend to favor local teams that you're all familiar with. These local teams might be good, but your reward for getting them right is to be in the crowd having to rely on how well you can pick the rest of the games. If you steer away from the local teams, there is a good chance you can distinguish yourself on that alone, and have a lot fewer co-leaders to overcome based on the non-regional team's games.

A good example of this happened in the 1999 tournament. Fans within the ACC region picked Duke to win it all, by an overwhelming margin. Only a small number went with the very powerful Connecticut and it paid off for them. Many pool winners who won because of Connecticut actually didn't do that well in the rest of their picks. But had Duke won, those that picked Duke couldn't have gotten away with an otherwise weak bracket.

7. Put historic stats into perspective.

The web is full of advice with all sorts of amazing stats. Some note the fact that only once since 1985 has a team repeated as national champion. Others might note how common it is for at least one 12th seed to win in the first round. But these stats need to be put into perspective. Yes it is hard for a team to repeat as national champion. But don't forget, that most of this phenomenom stems from the fact that it is hard to win the national championship. If 1 team has done it in the last 27 attempts, that is a 3.7% chance of a repeat. Of the 68 teams in the tournament, on average each team has about a 1.5% chance of winning. So the defending champion on average does better than most teams. Detailed analysis shows that the defending champion does do slightly worse than expected based on seeding, but this deserves to be put carefully into context.

8. Be very cautious with "common" upset picks

There is a very good chance that, each year, at least one 12th seed will beat a 5th seed. But what does that mean to you? If you pick one 12th seed to win, chances are you're going to pick the wrong one. Some other 12th seed will be the one to win and you will get two games wrong instead of one. Unless you have studied the teams and can identify a 12 seed that is particularly strong and/or a 5th seed that is particularly weak, you should not haphazardly pick such an upset.

Unless your pool gives points for picking the same number of 5th seeds that actually will advance, all of this advice based on the average number of a seed advancing is BAD ADVICE.

9. Don't forget that no one is invincible.

This is particularly true in the semi-finals and finals. Any team that gets that far can be assumed to be competitive in upcoming games. Historically since 1985, 42% of the first seeds have made it to the semifinals. In the semi-final game the first seed wins 56% of the time (keep in mind that some cases they are playing against another first seed). The second seed wins 50% of the time and the third seed wins 64% of the time. All other seeds combined win 32% of the time.

In years in which there is an overwhelming favorite, you can distinguish yourself from the crowd by picking the second most likely team. Generally, before the last round only two entrants will have a chance to win the pool. You want to be one of them. If you pick a strong team that not many others pick, you have a better chance of being in the battle than if you pick the same team as many others.

10. When in doubt, lean toward the major conferences.

The stength of each conference varies from year to year, but it is easy for a very good team to fly below radar throughout the season if they are competing in a very powerful conference. Furthermore, these teams are more battle tested than teams that might be very good, yet playing in a weaker conference.

11. Don't hastily go below 3rd seed to make the semi-finals.

Almost every year someone below the 3rd seed makes the semifinals. But if you pick a team below the 3rd seed, chances are very good that you're going to pick the wrong team. There are 52 teams seeded 4th through 16th (not counting play-in losers). You get no points for guessing that one of the 52 will advance. You only get points for guessing which ones. That is very hard to do.

12. Do your homework.

Bottom line, there is much to be learned from online analysis. However, nothing is guaranteed. This is especially important in picking your semi-final and final winners. Look for teams who are hot. And look for teams with experience.

13. Look for "value" if your pool's scoring offers bonuses for picking upsets.

For example, if your scoring system offers a bonus of "seed difference times points per round" for correctly picking upsets, it is actually statistically more worthwhile to pick 10th, 11th, and 12th seeds in the first two round. This may vary depending on how many points are offerred per round. Usually, even when there are generous bonuses for picking upsets, it still is better to pick the #1 seeds.

14. Work extra hard for the "close" picks.

The 8th versus 9th seed game is almost a toss up. Historically the 9th seed has done a little better (52% to 48%), but even that edge has little statisical significance. The 7th seed has a 62% to 38% edge over the 10th seed. Very likely, those who correctly pick these games right will be on top after the first round. Those that pick a lot of upsets outside these games generally fall behind. Not that there won't be upsets, its just that not many will pick the right upsets. Or, those that do pick the right upsets also manage to pick a lot of wrong upsets too.

One interesting fact, even though the 9th seed has a slight edge over the 8th seed, when the 8th seed does manage to advance, they are 2.8 times more likley thant the 9th seed to knock off the top seed in the second round.

Create your own PickHoops office pool