There is always so much discussion about the early rounds heading into the tournament. Which upsets should you pick? Who is this year’s Cinderella, who makes a run to the Sweet 16 as an unexpected double-digit seed? Which higher-seeded team could be in danger of an early letdown?

Well, maybe we spend too much time worrying about early-round upsets and not enough time dissection potential Elite Eight and Final Four matchups. All of the talk about early rounds is deserved – the upsets are part of what makes March Madness so great – but if we’re talking about putting together the best bracket, let’s pay as much, if not more, attention at the later rounds.

TOURNAMENT RUN

Bracket busted? Build a
streak of correct picks. 

PLAY TOURNAMENT RUN
We took a look at the data from the past five years of the Capital One Bracket Challenge Game, the official bracket game of the NCAA tournament. There have been more than two million entries, but each year, there has been one outright winner.

Let’s take a look at how the five winners did, round by round.

Reminder: All rounds have equal value. So, correct picks in the traditional first round get one point, and that doubles up to getting 32 for the correct national champion.

  FIRST ROUND SECOND ROUND SWEET
16
ELITE
EIGHT
FINAL
FOUR
CHAMP. TOTAL
2015 25 30 28 32 32 32 179
2014 25 22 24 32 32 32 167
2013 22 24 28 24 32 32 162
2012 23 24 28 24 32 32 175
2011 25 20 20 16 32 32 145

Note: For years in which the Bracket Challenge Game included the First Four, that data was excluded for this exercise. It did not impact which bracket won the challenge.

Some important takeaways from the data:

Every one of the past five winners had the participants and winner of the national championship game right.

It doesn’t matter how many correct upsets you pick in the first round if you lose a championship game team. It’s hard to fill out a bracket from the championship game out, especially because matchups can be so important, but it may not be a bad idea to think about who can make it that far – and avoid picking those teams to be upset.

A few misses in the first round are OK. In fact, 25 was the best first-round score of the five winners.

Let’s look at the 2015 winner (pictured above), for recency’s sake and the fact that it was the best performer of the five winners. Even with the great score of 179, it missed the Georgia State-Baylor upset. It missed the UAB-Iowa State upset. It picked Harvard and Eastern Washington, but missed those.

The key? The missed picks turned out to be pretty inconsequential. In general, it’s better to miss an upset than to pick one, be wrong, and have that winning team go on another two or three rounds – or even worse, to the Final Four. Which leads us to …

When you're wrong, be wrong with everyone.

Basically, there’s not much of a need to find the one crazy early upset that no one else is going to get. If there’s a George Mason-esque run to the Final Four, you’ll still be in good shape – the chances of someone else having that and the other three Final Four teams are slim.

Of the past five winners, three scored 25 in the first round. Incredibly, for the 2011 winner, that was the best round until the Final Four. This is closely related to the lesson described above, as it goes to show how much more important getting later rounds right is than early rounds. Now, 2011 was the most unpredictable tournament in recent memory, so those kinds of numbers wouldn’t hold up in other years like they did then. The key for the 2011 bracket was that where it went with surprises, it was right – all the way up to UConn beating Butler in the championship. Where it was wrong (like missing VCU in the Final Four), almost everyone else was wrong, too.

Is there a fool-proof way to pick a bracket? No. But there are definite trends from previous Bracket Challenge winners that we can glean some lessons from. When you’re filling out your bracket this year, remember: the later the round, the more important it is to be right.