On Monday, my colleague Carl Bialik looked at the odds of taking $1 billion out of Warren Buffett’s wallet by completing a perfect NCAA tournament bracket.
They’re not good.
The simplest way to estimate your chances is by assuming that the outcome of each game is equivalent to a coin flip. There are 63 games in the tournament (the four “play-in” games aren’t included in Buffett’s challenge). What are the odds of having a fair coin come up heads 63 times in a row? Just one chance in 9,223,372,036,854,780,000 (one in 9.2 quintillion).
But as Carl explains, this simple assumption isn’t a smart one. Many games have a clear favorite, and your odds of picking the winner are much better than even. What if you used the FiveThirtyEight interactive bracket to make your picks, choosing the favorite in each game?
Rather than taking any statistical shortcuts, I calculated these odds for all 63 games. The model we use to create our bracket estimates the probability of any team defeating any other team in any given round.
The first tip-off scheduled for Thursday pits Ohio State against intrastate rival Dayton. Ohio State has a 75 percent chance of winning, per our model. The next game is a gimme: Wisconsin has a 93 percent chance of beating American University.
But it isn’t long before the road gets rougher. Cincinnati will play Harvard early on Thursday, and Harvard is much stronger than its #12 seed suggests. Cincinnati is a favorite, but only barely, with a 58 percent chance of winning.
Another game, scheduled for Thursday evening, really is the equivalent of a coin flip: We list Texas as a 50.1 percent “favorite” to beat Arizona State. (You can see your odds of surviving each game in the table that accompanies this post.)
You’ll have about a 1 in 85 chance of completing Thursday with a perfect bracket if you follow FiveThirtyEight’s recommendations.
Things get much worse from there.
As the tournament proceeds — and also-rans are eliminated from the field — the games become more competitive. On average, favorites in the first round have a 78 percent chance of winning. That probability drops to 68 percent for the second round and 61 percent for the Sweet Sixteen and beyond.
By the time the national championship game is played, your odds of winning Buffett’s money are just 1 in 7.4 billion.
So, you’re telling me there’s a chance? One chance in 7.4 billion is a lot better than one in 9.2 quintillion; it’s more than a billion times better. (CLICK-BAIT HEADLINE: HOW A BROOKLYN MOM BECAME A BILLION TIMES MORE LIKELY TO WIN WARREN BUFFETT’S MONEY.) It’s also toward the low end of the range of estimates that statisticians provided to Carl.
Odds are that Buffett’s bracket isn’t worth your time, however. Having a 1-in-7.4 billion chance of winning a billion dollars is worth the equivalent of 14 cents. That’s before accounting for taxes — or for the possibility that the prize pool will be split among multiple winners.
| Game | Date | Favorite | Underdog | Win % | Cumulative Probability |
|---|---|---|---|---|---|
| 1 | 3/20 | Ohio St. | Dayton | 75.3% | 1 in 1.3 |
| 2 | 3/20 | Wisconsin | American U. | 92.8% | 1 in 1.4 |
| 3 | 3/20 | Pittsburgh | Colorado | 72.4% | 1 in 2.0 |
| 4 | 3/20 | Cincinnati | Harvard | 58.0% | 1 in 3.4 |
| 5 | 3/20 | Syracuse | W. Michigan | 88.3% | 1 in 3.9 |
| 6 | 3/20 | Oregon | BYU | 64.7% | 1 in 6.0 |
| 7 | 3/20 | Florida | Albany / MSM | 98.8% | 1 in 6.0 |
| 8 | 3/20 | Michigan St. | Delaware | 91.2% | 1 in 6.6 |
| 9 | 3/20 | Connecticut | St. Joseph’s | 67.3% | 1 in 9.8 |
| 10 | 3/20 | Michigan | Wofford | 95.4% | 1 in 10.3 |
| 11 | 3/20 | Saint Louis | NCSU / Xavier | 57.7% | 1 in 18 |
| 12 | 3/20 | Oklahoma | N. Dakota St. | 63.8% | 1 in 28 |
| 13 | 3/20 | Villanova | Milwaukee | 94.7% | 1 in 30 |
| 14 | 3/20 | Texas | Arizona St. | 50.1% | 1 in 59 |
| 15 | 3/20 | Louisville | Manhattan | 93.1% | 1 in 63 |
| 16 | 3/20 | San Diego St. | New Mexico St. | 74.5% | 1 in 85 |
| 17 | 3/21 | Duke | Mercer | 92.9% | 1 in 91 |
| 18 | 3/21 | Baylor | Nebraska | 70.3% | 1 in 130 |
| 19 | 3/21 | New Mexico | Stanford | 63.8% | 1 in 204 |
| 20 | 3/21 | Arizona | Weber St. | 97.8% | 1 in 209 |
| 21 | 3/21 | Tenn. / Iowa | Massachusetts | 67.6% | 1 in 308 |
| 22 | 3/21 | Creighton | UL-Lafayette | 88.3% | 1 in 349 |
| 23 | 3/21 | Kansas | E. Kentucky | 92.4% | 1 in 378 |
| 24 | 3/21 | Oklahoma St. | Gonzaga | 52.0% | 1 in 726 |
| 25 | 3/21 | Memphis | George Wash. | 55.0% | 1 in 1,321 |
| 26 | 3/21 | Wichita St. | Cal Poly / Tx. So. | 97.9% | 1 in 1,349 |
| 27 | 3/21 | North Carolina | Providence | 68.0% | 1 in 1,985 |
| 28 | 3/21 | VCU | S.F. Austin | 76.4% | 1 in 2,598 |
| 29 | 3/21 | Virginia | Coast. Carolina | 96.4% | 1 in 2,695 |
| 30 | 3/21 | Kentucky | Kansas St. | 73.9% | 1 in 3,645 |
| 31 | 3/21 | Iowa St. | N.C. Central | 81.2% | 1 in 4,487 |
| 32 | 3/21 | UCLA | Tulsa | 87.1% | 1 in 5,153 |
| 33 | 3/22 | Villanova | Connecticut | 63.6% | 1 in 8,107 |
| 34 | 3/22 | Louisville | Saint Louis | 82.5% | 1 in 9,828 |
| 35 | 3/22 | Wisconsin | Oregon | 74.0% | 1 in 13,290 |
| 36 | 3/22 | San Diego St. | Oklahoma | 53.4% | 1 in 24,879 |
| 37 | 3/22 | Syracuse | Ohio St. | 51.0% | 1 in 48,817 |
| 38 | 3/22 | Florida | Pittsburgh | 82.0% | 1 in 59,519 |
| 39 | 3/22 | Michigan | Texas | 77.4% | 1 in 76,877 |
| 40 | 3/22 | Michigan St. | Cincinnati | 70.4% | 1 in 109,223 |
| 41 | 3/23 | Virginia | Memphis | 71.8% | 1 in 152,162 |
| 42 | 3/23 | Kansas | New Mexico | 68.9% | 1 in 220,941 |
| 43 | 3/23 | Iowa St. | North Carolina | 52.0% | 1 in 425,213 |
| 44 | 3/23 | UCLA | VCU | 67.5% | 1 in 629,867 |
| 45 | 3/23 | Duke | Tenn. / Iowa | 71.9% | 1 in 875,923 |
| 46 | 3/23 | Wichita St. | Kentucky | 55.2% | 1 in 1,586,448 |
| 47 | 3/23 | Creighton | Baylor | 56.4% | 1 in 2,812,735 |
| 48 | 3/23 | Arizona | Oklahoma St. | 73.6% | 1 in 3,819,148 |
| 49 | 3/27 | Florida | UCLA | 74.2% | 1 in 5,148,674 |
| 50 | 3/27 | Creighton | Wisconsin | 52.0% | 1 in 9,909,744 |
| 51 | 3/27 | Kansas | Syracuse | 62.7% | 1 in 15,802,666 |
| 52 | 3/27 | Arizona | San Diego St. | 72.7% | 1 in 21,722,064 |
| 53 | 3/28 | Villanova | Iowa St. | 61.6% | 1 in 35,283,661 |
| 54 | 3/28 | Louisville | Wichita St. | 67.3% | 1 in 52,420,618 |
| 55 | 3/28 | Michigan St. | Virginia | 52.1% | 1 in 100,626,657 |
| 56 | 3/28 | Duke | Michigan | 54.3% | 1 in 185,416,034 |
| 57 | 3/29 | Florida | Kansas | 58.9% | 1 in 314,845,275 |
| 58 | 3/29 | Arizona | Creighton | 69.2% | 1 in 454,867,222 |
| 59 | 3/30 | Michigan St. | Villanova | 54.6% | 1 in 832,622,741 |
| 60 | 3/30 | Louisville | Duke | 67.1% | 1 in 1,241,647,122 |
| 61 | 4/5 | Florida | Michigan St. | 57.6% | 1 in 2,153,920,866 |
| 62 | 4/5 | Louisville | Arizona | 54.4% | 1 in 3,956,466,819 |
| 63 | 4/7 | Louisville | Florida | 53.3% | 1 in 7,419,071,319 |
CORRECTION (March 18, 7:15 p.m.): My original version of this post listed incorrect opponents for games #47 and #48. These have now been corrected. However, my error also affected the overall odds of a perfect bracket (which were originally about 1 in 6 billion). Sorry, folks.
