Pete Tiernan started maintaining a database on NCAA tournament teams and results in 1991 in the vain hope of finding a way to win his office bracket pool. His research covers the tournaments since the field of teams was expanded to 64 in 1985. When he isn’t squirreled away crunching numbers on the tournament, Tiernan is a director of marketing for a simulation software company in Michigan. Here is his formula for success in finding winners in the NCAA bracket:
OVERVIEW
Fifty-five of the last 72 teams to reach the Final Four have been seeded No. 1 (32 times), No. 2 (15), or No. 3 (8). While fourth-seeded teams have reached the semifinals seven times, they’re not worth picking because they usually have to beat a top seed to advance–and there isn’t a set of conditions to predict such an upset reliably.
This means you should advance all 1 seeds to the Elite Eight, and move 2 and 3 seeds to the Sweet Sixteen, where they square off against each other. That leaves you with five matchups to predict in round one involving the four through 13 seeds. These matchups break down into two types: the toss-ups (7 vs. 10 and 8 vs. 9) and longshot games (4 vs. 13, 5 vs. 12 and 6 vs. 11) involving potential upsets.
FIRST-ROUND TIPS
If you fill out your bracket with the four top seeds in the Final Four, you should know that it has never happened before.
In the 8 vs. 9 matchup, 9 seeds have actually prevailed in 39 of the 72 games (a 54 percent winning rate). In 7 vs. 10 matchups, the seventh-seeded team has a 43-29 record.
In the longshot matchups, the higher seeds (4, 5 and 6) have beaten their lower seed opponents (13, 12, and 11 respectively) 72 percent of the time (155-61).
SECOND-ROUND TIPS
The toughest game to predict in round two is the 4 (or 13) vs. 5 (or 12) seed matchup. Four seeds have played 5 seeds 39 times in the last 18 tournaments and their record is a solid 24-15 (62 percent).
You can improve this rate to 30-9 (77 percent) by taking 5 seeds that: A) Score more than 77 points per game, B) Are on no better than a two-game winning streak and C) Get more than 17 percent of their scoring from the bench. Five seeds with these attributes are 8-1 against 4 seeds.
SWEET SIXTEEN
The Sweet Sixteen matchup of the 2 and 3 seeds is a key game to predict correctly. For one thing, 2s or 3s only win 49 of the 72 games (2s winning 34 times and 3s winning 15). More important, if you get this pick wrong, you create a false matchup in the semifinals that may cause you to continue advancing the wrong team.
If you play it safe and just pick the 2 seeds; you’d be right 34 of 72 times.
THE ELITE EIGHT
If you’re only advancing the top three seeds to the Elite Eight, you just need to worry about two possible matchups: 1 vs. 2 and 1 vs. 3 games. If you simply advanced 1 seeds, you’d be right in 32 of 72 games.
In 1 vs. 3 games, take top seeds unless the 3 seed: A) went to the previous year’s tournament, B) scores at least 77 points a game and wins by more than seven, C) gets more than 30 percent of its scoring from guards.
The 1 vs. 2 game is a little more complicated. For a 2 seed to advance, the team must: A) have appeared in at least four tourneys in a row, B) score more than 74 points a game and win by more than 10, and C) get no more than 55 percent of their scoring from guards.
FINAL FOUR
Under this system, there are only six potential matchups that could happen in the Final Four: 1 vs. 1, 1 vs. 2, 1 vs. 3, 2 vs. 2, 2 vs. 3 and 3 vs. 3. Of the 36 semifinal games played since 1985, 23 of them have involved these matchups.
In 2 vs. 3 and 2 vs. 2 matchups take the team with a higher percentage of scoring from its backcourt.
Take top seed in 1 vs. 3 games.
In 1 vs. 2 games, take the top seed unless the 2 seed scores more than 77 points per game and wins by more than 10.
In 1 vs. 1 games, eliminate any team that either doesn’t have a coach with at least three March Madness appearances or doesn’t score more than 75 points per game. If that’s a wash, take the 1 seed that gets a higher percentage of scoring from its guards.
