Nobody picks a perfect NCAA tournament bracket, so there is no perfect advice for filling one out. If we at The Ringer could predict how every wayward bounce will play out over the next three weeks, we’d pursue more profitable enterprises than dispensing tips to the masses.
But just because no advice is perfect doesn’t mean that all pieces of advice are equally flawed. March Madness has inspired a whole set of bracket tips and tricks, but only some of them pass statistical muster. Before completing your selections this week, learn which ideas to take to heart—and which to dismiss as unsubstantiated myth.
1. Don’t assume a superior defensive team is the best pick.
Throughout this piece, we’ll look at tournament results since 2002, the first year that the advanced statistics from KenPom were available. And the first myth that data disproves is that defense is more likely to win out in the NCAA tournament.
In 483 of the 1,112 games in this sample, one team rated better on offense while the other rated better on defense, using KenPom’s numbers. The team with the offensive advantage won 50.5 percent of those games; the team with the defensive advantage won 49.5 percent. Essentially identical! (For reference, the 629 teams that held the advantage on both sides of the ball won 80 percent of their games.)
2. Pick against teams with imbalanced offensive and defensive profiles.
When a team has a wide disparity between its offensive and defensive KenPom ratings, it is both more likely to underperform and less likely to overperform its expectation by seed than a team whose ratings are more balanced. (Expectation by seed here means how far a team would advance if no upsets occurred, e.g., a no. 1 seed would expect to reach the Final Four; a no. 2 would expect to reach the Elite 8; a no. 3 and no. 4 would expect to reach the Sweet 16; and so on.) Imbalanced teams are both more likely to lose in an upset and less likely to deliver an upset themselves.
Among teams seeded 1 through 8 in this year’s field, these teams have the largest differentials between offensive and defensive performance:
- VCU (no. 8 seed in East region): 172nd on offense, seventh on defense
- Kansas State (no. 4 seed in South region): 102nd on offense, fourth on defense
- Villanova (no. 6 seed in South region): 16th on offense, 73rd on defense
- LSU (no. 3 seed in East region): 10th on offense, 62nd on defense
- Wofford (no. 7 seed in Midwest region): 12th on offense, 63rd on defense
- Iowa State (no. 6 seed Midwest region): ninth on offense, 59th on defense
- Wisconsin (no. 5 seed in South region): 52nd on offense, third on defense
And here are the teams seeded ninth or lower with the most balanced profiles:
- Temple (no. 11 seed in East region): 86th on offense, 90th on defense
- Minnesota (no. 10 seed in East region): 54th on offense, 44th on defense
- Arizona State (no. 11 seed in West region): 68th on offense, 79th on defense
- Vermont (no. 13 seed in West region): 103rd on offense, 84th on defense
- Prairie View A&M (no. 16 seed in West region): 226th on offense, 205th on defense
- UCF (no. 9 seed in East region): 58th on offense, 36th on defense
- Seton Hall (no. 10 seed in Midwest region): 74th on offense, 49th on defense
Of course, nobody should pick Prairie View A&M over Gonzaga. But UCF is balanced, while its first-round opponent, VCU, is not; the same goes for Wofford and Seton Hall. Perhaps start with this guiding principle when looking for an upset, even if two mid-major darlings are victims in the process.
3. Playing at a slow pace doesn’t make a favored team more susceptible to an upset.
Virginia’s historic 1 vs. 16 loss last season lent credence to the myth that a slower tempo makes a lesser team more likely to keep the score close, and thus capable of springing a surprise. To test the theory, we split all of the top four seeds since 2002—whose losses produce the most memorable and unexpected upsets—into three groups based on pace: those that ranked in the top third in the country in tempo, those that ranked in the middle third, and those that ranked in the bottom third. Then we tested to see how often teams in each group were upset in the first round.
Upset Frequency by Pace
| Pace Grouping | Total Teams | First-Round Upsets | Upset % |
|---|---|---|---|
| Pace Grouping | Total Teams | First-Round Upsets | Upset % |
| Top Third | 108 | 10 | 9.3% |
| Middle Third | 84 | 10 | 11.9% |
| Bottom Third | 80 | 7 | 8.8% |
There’s no signal whatsoever—so don’t discriminate against Virginia, Michigan, Virginia Tech, and Kansas State, all of whom ranked below 300th nationally in pace this year, solely because they plod.
4. 11-over-6 is the best place to look for a first-round upset.
The 5-over-12 upset is the perennial sexy choice, but the real action comes one rung down the seed line, where teams like Loyola-Chicago last year, VCU in 2011, and George Mason in 2006 all started their memorable Final Four runs. Since 2002, 11-seeds have pulled the most first-round upsets—even more than 9s, who should be almost indecipherable from their eighth-seeded opponents.
:format(webp):no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/15968959/Frequency_of_First_Round_Upsets_by_Seed__2002_18.png)
5. Stay away from picking 8- and 9-seeds to upset a no. 1.
This tantalizing option is a problem for two reasons. First, even picking the winner of the 8-9 game is a difficult bet. No. 8 seeds have won just 57 percent of their first-round contests since 2002, so based on this historical precedent, the 8- or 9-seed in question has only about a coin-flip chance to win one game, let alone two. And even once they have advanced to the second round, they’re unlikely to complete the upset. Top seeds have won a combined 85 percent of their games against their 8 or 9 counterparts.
Here’s the overall breakdown of teams to qualify for the Sweet 16 since 2002. Notice that more teams seeded 10th, 11th, and 12th have reached the second week than teams seeded eighth or ninth.
:format(webp):no_upscale()/cdn.vox-cdn.com/uploads/chorus_asset/file/15968958/Sweet_16_Seed_Distribution__2002_18.png)
No. 8 Syracuse is sure to be an enticing choice this year because of the Orange’s tournament history compared with that of Gonzaga, the no. 1 team in the West region. This history, though, suggests it’s far safer to scoot Gonzaga to the Sweet 16.
6. Take a flier on a First Four team to make a run.
No. 11 seeds that must win a First Four game to reach the 64-team tournament field have historically been more likely to advance in every round than no. 11 seeds that don’t have to qualify via the play-in games in Dayton.
11-Seed Tournament Journeys
| Type of Team | Won a Game | Reached Sweet 16 | Reached Elite Eight | Reached Final Four |
|---|---|---|---|---|
| Type of Team | Won a Game | Reached Sweet 16 | Reached Elite Eight | Reached Final Four |
| First Four 11s | 55% | 27% | 9% | 9% |
| Non–First Four 11s | 42% | 18% | 7% | 4% |
So beware, 6-seeds Maryland and Buffalo—but also beware, LSU and Texas Tech, the no. 3 seeds in this year’s regions that host an 11 vs. 11 play-in game, and the teams that might have to beat a lesser-ranked team full of momentum to reach the Sweet 16.
7. Don’t depend on a ball-dominant star to lead a team to glory.
From Kemba Walker and Shabazz Napier leading UConn to championships, to the likes of Frank Kaminsky (Wisconsin) and Trey Burke (Michigan) sparking their teams’ runs to the title game, the tournament is rife with examples of players, typically guards, who shepherd their team to later rounds by virtue of their singular dominance. It’s a tricky thing to calculate statistically, but we can get an approximate count by defining a team as having a ball-dominant star if a single player leads his team in both points and assists per game; that combination generally means the player can create the best chances for both himself and others.
But among teams seeded 1 through 8—in other words, first-round favorites—23 percent of teams with a ball-dominant player since 2002 overperformed their expectation by seed. What’s that ratio for teams without a ball-dominant star? Twenty-two percent. Basically the same. So as tempting as the dream of Ja Morant propelling Murray State on a lengthy run might be or Carsen Edwards pushing Purdue to the Final Four, don’t make those picks on such wishes alone.
8. Ignore the seeds when no. 1s play no. 2s.
In 30 matchups of a no. 1 vs. no. 2 seed since 2002, the two groups have played to an even 15-15 split. In matchups occurring only in the Elite Eight, the no. 2 seeds have actually been better, winning 12 of 22 contests.
The no. 1 seeds have been more dominant in the games they’ve won, with an average margin of victory of 12 points, versus just seven points the other way. But the game results speak for themselves, so don’t hesitate to pick a 2-over-1 upset in this year’s Elite Eight. The options are there: The top eight teams by KenPom are all seeded no. 1 or 2, so the talent pool is roughly similar across the top two seed lines.
9. Be more adventurous with your upset picks in later rounds.
Once upon a time, the smartest brackets pushed three no. 1 seeds to the Final Four along with a no. 2 or maybe a 3 to round out the quartet. No longer. From 2002 through 2009, here were the worst-seeded teams to reach the Final Four by year: 5, 3, 3, 5, 11, 2, 1, 3. Here’s that list since 2010: 5, 11, 4, 9, 8, 7, 10, 7, 11. That second group includes a lot more higher numbers.
The average sum of the seeds of the teams to reach the later rounds has grown this decade compared to last, too, meaning more and more lower-seeded teams are advancing farther and farther in the tournament.
Later Rounds’ Changing Composition
| Year Range | Sweet 16 | Elite Eight | Final Four |
|---|---|---|---|
| Year Range | Sweet 16 | Elite Eight | Final Four |
| 2002-09 | 66.0 | 23.1 | 9.3 |
| 2010-18 | 75.4 | 28.9 | 15.2 |
10. But don’t be adventurous with the pick for overall champion.
Well, they’re advancing farther and farther until a very specific limit. In other words, embrace the upsets up until the final round. Besides the seventh-seeded, Napier–led Connecticut team in 2014, every champion since 2002 has been a top-three seed: 11 no. 1s, two no. 2s, and three no. 3s. Other teams have come close (2014 Kentucky and 2011 Butler were both 8-seeds; 2002 Indiana and 2010 Butler were both 5s), but no other surprise team has sealed the deal.
So enjoy the outlandish up to a certain point. And then pick Duke or North Carolina or Michigan State like always. It’s just the way most tournaments work.
https://www.theringer.com/2019/3/18/18270554/ncaa-tournament-bracket-tips
2019-03-18 10:20:00Z
52780242560977
Tidak ada komentar:
Posting Komentar