Every March, roughly 100 million Americans do something mathematically irrational: they fill out a bracket. They pick 63 winners across six rounds of the NCAA Men's Basketball Tournament, enter their predictions into an office pool, and wait. Most brackets are busted within 48 hours. All brackets are busted within three weeks. And yet, every year, the ritual repeats — because somewhere in the back of every bracket-filler's mind lives the same impossible dream.

What if this is the year?

It won't be. The math is extraordinarily clear about that. But the math is also fascinating, and so are the stories it produces: a neuropsychologist who didn't watch a single game for two days and set the all-time record; a billionaire who offered a thousand million dollars just to see if anyone could do it; and a tournament structure that, year after year, produces just enough chaos to keep us all believing.

The Odds: A Number Too Large to Imagine

Start with the simplest version of the math. There are 63 games in the bracket (excluding the First Four play-in games, which virtually no pool requires you to pick). Each game has two possible outcomes. That means the total number of possible brackets is 2 raised to the 63rd power — written out, that's 9,223,372,036,854,775,808. Nine point two quintillion. One quintillion is one billion billions.

9.2Q Possible brackets
(random picks)
120B Possible brackets
(knowledgeable fan)
63 Games to predict
correctly
366 Years until expected
perfect bracket*

*If every American filled out a unique, expert-level bracket every year

The coin-flip calculation isn't entirely realistic, of course — No. 1 seeds beat No. 16 seeds nearly every time, which is useful information. Apply a more generous model: assume a knowledgeable fan picks each game correctly 66.7% of the time, the estimate used by the NCAA's own Bracket Challenge data. The odds improve dramatically, falling to roughly 1 in 120.2 billion. Better — but still essentially zero. To put both numbers in tactile perspective: there are approximately 7.5 quintillion grains of sand on Earth. Randomly picking the same grain as a stranger is more likely than a coin-flip perfect bracket. And winning the Powerball jackpot is roughly 2,400 times more likely than the knowledgeable-fan perfect bracket.

"If every person in the United States filled out a completely unique, expert-level bracket, we'd expect to see a perfect bracket in about 366 years — around the year 2385."

— NCAA.com, Bracket IQ

Even in the age of artificial intelligence, no computer model has cracked it. The bracket's resistance to prediction isn't a bug — it's the whole point. Basketball is what analysts call a "make-or-miss" sport. The best shooters make around half their twos and 40% of their threes. Sixty-three consecutive games, each injecting its own variance, compound into near-infinite unpredictability. The madness, it turns out, is structural.

Who Came Closest? The Man Who Didn't Watch a Game

In the entire history of the NCAA Tournament — spanning more than 40 years of online bracket tracking and estimates of over 3 billion brackets filled out in total — no one has ever achieved a perfect bracket. The closest anyone has officially come was 49 games. And the story of how it happened is perfectly, absurdly human.

Gregg Nigl — 49 Correct Picks, 2019

Gregg Nigl is a neuropsychologist from Columbus, Ohio. On the first Thursday of the 2019 tournament, he called in sick from work. He was genuinely unwell — so unwell that he never turned on the television. He didn't watch a single game.

He'd filled out four brackets that year, including one called "center road" entered into a small group of friends on NCAA.com's Bracket Challenge. He almost didn't fill that one out at all. On Saturday he felt better, packed his family into the car for a 12-hour drive to Vermont, and started listening to games on satellite radio. He still hadn't checked his bracket.

On Monday, NCAA.com called him. They had been tracking tens of millions of brackets across every major platform — ESPN, CBS, Yahoo, Fox, Sports Illustrated. Of all of them, one remained perfect: "center road." Nigl had no idea. "I thought it was a joke," he said.

His bracket had correctly predicted all 48 games through two full rounds. It survived Sunday's scares — Duke's narrow escape over UCF, Tennessee's overtime win — and entered the Sweet 16 as the only perfect bracket in history. It finally fell on game 50, when No. 3 seed Purdue beat No. 2 seed Tennessee 99–94 in overtime. Not even a wild upset — just the randomness of two evenly matched teams. Buick flew him to Anaheim for the remaining games. He did national media interviews. His other brackets, it turned out, had only one Final Four team correct.

Previous record: 39 consecutive correct picks, achieved in 2017. Nigl shattered it by 10 games.

Since Nigl's run, no bracket has come close. In 2023, the last perfect bracket was busted on game 25 — when 16-seed Fairleigh Dickinson knocked off 1-seed Purdue. In 2025, the final perfect bracket fell on game 31. The record of 49 stands, and given the mathematics, it may stand for a very long time.

The Billion-Dollar Prize (and the Man Who Offered It)

If someone did somehow produce a perfect bracket, what would it be worth? For one year, the answer was a billion dollars.

In January 2014, Quicken Loans and Warren Buffett's Berkshire Hathaway announced the "Billion Dollar Bracket Challenge" — open to any U.S. adult who could correctly predict all 63 games of the tournament. The prize: 40 annual installments of $25 million, or a lump sum of $500 million. Buffett, whose Berkshire Hathaway was insuring the prize, understood the math perfectly. At 9.2 quintillion-to-one odds, it was less a contest than a very public demonstration of probability. Nobody won. Nobody was close.

The Buffett Timeline

2014: $1 billion prize to the public for a perfect 63-game bracket — nobody claimed it. 2016: Contest moved internal to Berkshire Hathaway employees (~375,000 people); prize reduced to $1 million/year for life for a perfect Sweet 16. Nobody won. 2017–2024: Rules relaxed repeatedly; still no winner of the grand prize. 2025: Buffett, age 94, simplified further — correctly pick 30 of the first 32 games to win $1 million. "I'm getting older," he told the Wall Street Journal. "I want to give away a million dollars to somebody while I'm still around." Finally, in 2025, a winner emerged — keeping their identity private. Eleven runners-up each received $100,000.

The Buffett saga neatly illustrates the bracket's strange cultural power. A billionaire repeatedly lowered the bar over eleven years, trying to give away money he was mathematically certain would never be claimed under the original rules. The 2025 winner needed to get 30 of 32 first-round games right — impressive, but nowhere near the original 63-game requirement. The true $1 billion prize remains unclaimed, and will almost certainly remain so forever.

— ✦ —

By the Seeds: What 40 Years of Data Actually Shows

The tournament's format — 64 teams seeded 1 through 16 in each of four regions — creates a rich historical record. Since the bracket expanded to its modern 64-team format in 1985, we have 40 complete tournaments to analyze. The NCAA's own records reveal a picture of both order and chaos: the favorites win most of the time, but the exceptions are vivid enough to haunt every bracket.

First-Round Win Records by Matchup — 1985 to 2025 (NCAA.com)

Matchup Record (Higher Seed) Win Rate Notes
1 vs. 16 158–2
98.8%
UMBC '18, FDU '23 Only 2 upsets ever
2 vs. 15 149–11
93.1%
3 vs. 14 137–23
85.6%
4 vs. 13 127–33
79.4%
5 vs. 12 103–57
64.4%
The classic upset pick 12 wins ~36%
6 vs. 11 98–62
61.3%
7 vs. 10 97–62
61.0%
8 vs. 9 77–83
48.1%
9-seed leads historically

*No. 7 row excludes one 2021 no-contest (Oregon/VCU, COVID protocols). Source: NCAA.com

The Only Two Times It's Ever Happened

UMBC def. Virginia — 2018

FDU def. Purdue — 2023

The 8 vs. 9 matchup is the bracket's great equalizer. These two teams are so closely matched that the committee essentially can't distinguish between them — and 40 years of results confirm it, with 9-seeds holding a slim 83–77 all-time edge. Pick either one. The 5 vs. 12 game is the other perennial topic: the 12-seed wins more than a third of the time, and in five separate tournaments (2002, 2009, 2013, 2014, and 2019) all three available 12-seeds pulled off first-round upsets. If you're not picking at least one 12-over-5, history suggests you're leaving points on the table.

Overall Tournament Records & Best Finishes by Seed — 1985 to 2025 (NCAA.com)

Seed Overall W–L Final Four Champ. Game Titles Note
1534–134664126
2373–15532125
3294–15617124
4250–1581542
5183–160940Never won
6168–159321Kansas '88
7141–159311UConn '14
8113–159641Villanova '85
998–160200
1097–159100Syracuse '16
11105–160600
1281–160000
1339–160000
1425–160000
1516–160000
162–160000

Final Four and Champ Game columns reflect total appearances since 1985. Source: NCAA.com

A few things jump from the data. Seeds 1 through 6 are the only ones with overall winning records — every seed from 7 onward loses more games than it wins across the full tournament. The No. 1 seed is the dominant force: 66 Final Four appearances, 26 championships, and a 78% overall win rate across all games. The next closest, No. 2 seeds, have won just five titles despite 32 Final Four appearances.

The strangest entry in the table is the No. 5 seed. Despite reaching four championship games and nine Final Fours, a No. 5 seed has never won the national championship — the only top-eight seed with that distinction. San Diego State in 2023 was the most recent to come close, falling in the title game.

The No. 11 seed deserves its own footnote. Of all the Cinderella seeds, the 11 has been the most persistently dramatic. Six No. 11 seeds have reached the Final Four: LSU in 1986, George Mason in 2006, VCU in 2011, Loyola-Chicago in 2018, UCLA in 2021, and NC State in 2024. Yet not a single one has made it to the championship game. They arrive as the tournament's most improbable heroes — then vanish, right before the finale, every time.

— ✦ —

Why We Keep Trying

The math is clear. The record is 49 games. The billion-dollar prize went unclaimed for over a decade, until the rules were simplified almost beyond recognition. And yet, every March, a hundred million people fill out brackets. They pick 12-seeds over 5-seeds. They agonize over 8 vs. 9. They pencil in a Final Four team from the 11-line and feel quietly good about it.

Gregg Nigl — the closest anyone has officially come — filled out his record-breaking bracket as an afterthought, in a pool he nearly didn't enter, while sick on a couch in Columbus. He wasn't even watching the games when history was being made. There's something instructive in that: the perfect bracket, if it ever happens, probably won't happen to someone who deserves it.

That's the tournament. That's the madness. No amount of data, however precisely sourced and carefully verified, fully insulates you from it. Pick your upsets. Back your Cinderellas. And accept that somewhere, someone filling out their bracket in five minutes on a Tuesday morning might — just might — get further than any computer model ever has.