Checkmate

It started off so well. 29 of 32 in Round 1 with all of the Sweet 16 still intact. Then Wisconsin knocked out my national champion pick and my entire bracket is worthless. So, what went wrong?

  • Don’t ignore the immeasurables. I was so immersed in the numbers from this season in the week leading up to the tournament that I completely ignored the factors that are either not quantifiable or difficult to objectively quantify. Villanova was the defending champ, and thus had a high likelihood of complacency in this year’s tournament. Unless a defending champion is returning an absurd amount of talent (like Florida in the 2007 tournament or Duke in the 1992 tournament), they are unlikely to pull off the repeat. In fact, history shows they’re much more likely to have an early exit than to make a deep run. Wisconsin, on the other hand, was wildly under-seeded and likely ticked off about it. They were set out to show the committee that they were wrong, and did so emphatically by handling Virginia Tech with relative ease in Round 1, followed by the biggest upset of the tournament in Round 2 by knocking off the defending champs.
  • Protect the king! As I mentioned in my Bracket Guidelines last week, the early rounds don’t mean a thing if your champion gets knocked out early. It is vitally important to pick a viable champion; a safe champion. Villanova, despite what my formula proclaimed about their regular season successes, was not a safe pick. Their roster wasn’t nearly strong enough to justify a repeat championship, and their coach’s tournament history outside of last year is very underwhelming. Last year’s run was a historical anomaly for Jay Wright; this year was par for the course.

Aside from the wacky East region, the rest of my bracket turned out just fine, with 12 of my Sweet 16 advancing and 6 of the Elite 8 remaining. That one missing piece though is an absolute crater. I’ll be sure to make mention of the non-quantifiable factors next year prior to the start of the tournament.

I may post a tournament recap in a couple weeks if there is an interesting conclusion but otherwise will be shifting into off-season mode, which means a much lower frequency of posts. I do have a few off-season features in the works that aim to answer some misunderstood statistical claims:

  • Is it really that difficult to beat a team for a 3rd time in a season?
  • Just how significant is home-court advantage, and how much of it is attributable to officiating deference to the home team?

I’ll also update standings on my Best Dual Sports schools and shift data tables (such as the Best Dual Sports and college football win probabilities) to a Google Sheets format going forward. Follow me on Twitter for notifications on my latest posts.

 

Advertisements

My 2017 Bracket Picks

My bracket, along with explanations of upset picks and Final Four games:

2017 Bracket

Round 1

UNC-Wilmington over Virginia: Regular season champion with a strong record away from home as well as in close games against one of the slowest paced teams in the field. The pace will keep it close enough for UNCW to have a chance to eke out a win.

USC over SMU: The play-in game is often more beneficial to the participant than the awaiting team for the first round, as there has been a play-in game winner every year since the field expanded to 68 in 2011. I like USC more than Kansas State as this year’s version.

Wichita St over Dayton: Vegas doesn’t even consider this an upset. Horrible seeding by the committee gives Dayton a brutal round 1 draw.

Middle Tennessee over Minnesota: Again, Vegas doesn’t see this as an upset. This year’s version of MTSU is even better than the one that knocked off 2 seed Michigan State last year

Rhode Island over Creighton: Rhode Island is one of the hottest teams in the field coming in.

Vermont over Purdue: Vermont has to be considered the hottest team in the field coming in, as they haven’t lost a game since December.

Michigan St over Miami FL: 9 over 8 isn’t much of an upset, but it’s really hard to see an Izzo team getting bounced in round 1 two years in a row. They’ll find a way to get the W.

 

Round 2

Michigan over Louisville: This year’s greatest story, winning 4 games in 4 days after a plane crash. Their run continues just a little longer as they avenge their 2013 title game defeat against the Cardinals.

Sweet 16:

Michigan over Oregon: Oregon can only go so far without Chris Boucher. His loss will be felt in a big way in this match-up.

Elite 8:

Arizona over Gonzaga: Avenging an early-season loss, the Wildcats were identified in my ratings as one of the four strongest teams this season.

Kentucky over North Carolina: Another upset featuring a rematch, but this time the same team advances. North Carolina’s resume was a bit inflated due to fortunate scheduling in conference play, and this was validated in a disappointing ACC tournament showing. Meanwhile, Kentucky went out of their way to play everyone in the non-conference, and while they didn’t win all of them they are certainly battle-tested despite cruising through the SEC this year.

Final Four:

Villanova over Arizona: Last year’s champion is back with very strong team. It’ll be tough in a pseudo road game, but the extensive tournament experience will pay off on the biggest stage.

Kansas over Kentucky: Talent vs. Talent. There’s just a little bit more on the Kansas side this year.

Title game:

Villanova over Kansas: My ratings identify these as the two best teams this year, with Villanova receiving the edge as the top team. Will this be vindication for BoW after another bad score in the Bracket Matrix? We’ll find out in about 3 weeks.

Enjoy the madness, everyone!

 

 

Bracket Guidelines

I thought about writing a full preview of the bracket, complete with match-up breakdowns and identifying the most likely contenders. Upon researching however, I found that everything I planned to include is already out there for consumption; I would not be adding any value by piling on with yet another opinion. Instead, the goal of this preview is to provide a guide on how to set up your bracket to put it in position to win an office pool. Will this guarantee that your bracket will win? Absolutely not. The idea is to get out of your own way and focus on the important aspect of the bracket: identify teams which are most likely to advance. I have a few rules I like to keep in mind when filling out a bracket, and I only go against them when I have a really good reason for it.

  • First, set up your bracket with the 1 seeds advancing to the Sweet 16 and the 2,3, and 4 seeds advancing to round 2. Historically, these seeds have had at least an 80% success rate in these games, and there is a considerable drop-off from the 4 seed to 5 seed in round 1. If you have a good reason to go against a couple of these, go for it. Otherwise leave these as they are and move on.
  • Protect the king. Think of each round of the bracket like a group of chess pieces. The king is the most important; you must protect it at all costs. This is your eventual champion, and it is important to not voluntarily drop a potential champion in the first weekend. Historically, there is a 91% chance that the eventual champion is a 1,2,3 or 4 seed, which is why the first rule is to keep all of them at least one round. Similarly, think of all the latest rounds as the more valuable chess pieces and the 1-win teams as pawns. Upsets are fun to pick, but if you only have that team getting to the round of 32 at the expense of a team that could make a deep run it isn’t worth it.
  • A coach shouldn’t be picked to go more than 2 rounds further than their current best round. For example, if a coach has never taken a team to a Final Four, his current team should not be picked as a national champion. Similarly, if this is a coach’s first tournament appearance, his team should not be picked beyond the 2nd round (ignore play-in games). This is not a cover-all-scenarios rule; there certainly have been exceptions. But in general, it’s a good idea to keep a coach’s track record in tournaments in mind.
  • Be aware of recent injuries to key players. If a team has gone through most/all of the regular season with injuries, then this can be ignored as we have a good sense of what that team is already. A player who has been sidelined since early February or later should be factored in.
  • Superstar players can take over a game. Look at the list of Naismith semifinalists and all-conference 1st team players from the power conferences. Teams with multiple players on these lists are more likely to advance than teams with no players on these lists.
  • Veteran players with tournament experience. Teams with a lot of players who have played in tournament games will tend to perform more consistently to or even above their expectations. The first-timers are more likely to falter under the brighter lights.
  • High win percentage away from home. Teams that struggled on the road or at neutral sites during the season won’t all of a sudden figure it out at tournament time.
  • Capability to blow out teams. If a team struggled to put away bad teams during the season, they are ripe for an upset in the tournament. Look for upsets of teams who weren’t able to crush anybody during the regular season.
  • Strong record in close games. The tournament is full of them, so pick teams who tend to fare well in a tight match-up.

Metrics are important, but I’d tend to shy away from them unless they have a proven success rate over time. If you have something that you think is time-proof, test it out on the Algebracket over multiple years.

I’ll post my own bracket as soon as it is complete. Good luck out there!

 

 

2017 Bubble Analysis

The bubble is big and soft this year. Lots of teams remain in contention, many of which have left quite a bit to be desired in their play thus far. Let’s take a look at the multi-bid conferences to sort out this mess before championship week hits full stride.

ACC: 6 Locks (North Carolina, Louisville, Florida St, Duke, Virginia, Notre Dame); 4 on the bubble (Miami FL, Virginia Tech, Wake Forest, Syracuse); Projection: 9 teams in (one of Wake Forest/Syracuse may be left out, unless both make deep runs in the ACC tournament)

American: 2 Locks (Cincinnati, SMU); 1 on the bubble (Houston); Projection: 2 teams in

A10: 2 Locks (Dayton, VCU); 1 on the bubble (Rhode Island); Projection: 2 teams in

Big East: 3 Locks (Villanova, Butler, Creighton); 4 on the bubble (Marquette, Seton Hall, Xavier, Providence); Projection: 7 teams in. Xavier can’t lose to DePaul in the Big East opening round. Seton Hall and Marquette square off and as long as it’s not a blowout both are safe, and Providence just needs to keep it competitive with Creighton to be safe.

Big Ten: 4 Locks (Purdue, Minnesota, Maryland, Wisconsin); 5 on the bubble (Northwestern, Michigan, Michigan St, Illinois, Iowa); Projection: 8 teams in. The loser of Michigan/Illinois on Thursday might need some outside help to get a bid (see: Conference USA below); they’ll be cutting it awfully close.

Big XII: 5 Locks (Kansas, Baylor, West Virginia, Iowa St, Oklahoma St); 1 on the bubble (Kansas St); Projection: 5 teams in. Kansas St went 8-10 in conference play; they’ll need at least two wins in the Big XII tournament to earn an at-large bid. They open with Baylor, then would play West Virginia.

Conference USA: 0 Locks; 1 on the bubble (Middle Tennessee); Projection: 1 team in. Middle Tennessee needs to earn their auto-bid to truly feel safe. They should be good for an at-large at 27 wins (and counting) and their 15 over 2 upset win last year, but the CUSA doesn’t have a strong reputation like the Missouri Valley or West Coast so this is one to keep an eye on.

Missouri Valley: 1 Lock (Wichita St); 1 on the bubble (Illinois St); Projection: 2 teams in. Illinois St won 27 games this year. Even with a weak schedule, that should be good enough out of a mid-major conference with a stellar reputation.

PAC12: 4 Locks (Oregon, UCLA, Arizona, USC); 1 on the bubble (California); Projection: 4 teams in. The top-heavy PAC12 will demand more wins than Cal came up with this year to secure an at-large bid.

SEC: 4 Locks (Kentucky, Florida, Arkansas, South Carolina); 3 on the bubble (Vanderbilt, Georgia, Ole Miss); Projection: 4 teams in. Unless Vanderbilt earns the auto-bid, they will have 15 losses on the year, and no team has earned an at-large with more than 14.

West Coast: 2 Locks (Gonzaga, St Mary’s) 1 on the bubble (BYU); Projection: 2 teams in. If BYU can pull an upset Monday night over St Mary’s they’ll have a shot at an at-large, and if they can grab another win Tuesday they’ll have earned the auto-bid.

Total: 33 out of 47 (11 auto plus 36 at-large) are secured with 13 more projected in for a total of 46/47 spots. The ACC and Big Ten tournaments could produce a lot of chaos this week, with each conference having multiple teams on the wrong side of the bubble that could play their way in. Middle Tennessee and Vanderbilt will also be interesting to follow.

 

Bracket Assessment: 10 Days left

We are starting to gain some clarity on which teams will be in the field. Some conference tournaments have begun and auto-bids will begin to be claimed this weekend.

Tournament Locks

These teams are in the tournament, regardless of what happens the rest of the way

1 Villanova Big East
2 Kansas Big 12
3 Gonzaga West Coast
4 Kentucky Southeastern
5 Oregon Pacific-12
6 Butler Big East
7 UCLA Pacific-12
8 Florida Southeastern
9 Arizona Pacific-12
10 Louisville Atlantic Coast
11 North Carolina Atlantic Coast
12 Baylor Big 12
13 Purdue Big Ten
14 SMU American
15 Duke Atlantic Coast
16 Cincinnati American
17 Florida St Atlantic Coast
18 St Mary’s CA West Coast
19 Creighton Big East
20 Minnesota Big Ten
21 West Virginia Big 12
22 Virginia Atlantic Coast
23 Wisconsin Big Ten
24 Wichita St Missouri Valley
25 Maryland Big Ten
26 Notre Dame Atlantic Coast
27 Iowa St Big 12
28 Oklahoma St Big 12
29 South Carolina Southeastern
30 Dayton Atlantic 10
31 Arkansas Southeastern
32 MTSU Conference USA
33 VA Commonwealth Atlantic 10

1 Win Away

These teams are super close to punching their ticket, one more win should do it. Bid stealing may become a concern without getting another win

34 Illinois St Missouri Valley
35 Virginia Tech Atlantic Coast
36 Miami FL Atlantic Coast
37 Xavier Big East
38 Michigan Big Ten
39 Michigan St Big Ten
40 Nevada Mountain West
41 USC Pacific-12
42 Northwestern Big Ten

Remaining Contenders

These teams can still get an at-large bid, but they’ll need a few more wins and hope for minimal bid stealing in conference tournaments. The teams further down this list need to make the deepest conference tournament runs to secure a bid, and those not in a power conference very likely need to earn an auto bid to truly feel safe.

43 Seton Hall Big East
44 Providence Big East
45 Houston American
46 Marquette Big East
47 California Pacific-12
48 Illinois Big Ten
49 Valparaiso Horizon
50 Vanderbilt Southeastern
51 Rhode Island Atlantic 10
52 Georgia Southeastern
53 Kansas St Big 12
54 Mississippi Southeastern
55 UNC Wilmington Colonial
56 Ohio St Big Ten
57 BYU West Coast
58 Colorado St Mountain West
59 Wake Forest Atlantic Coast
60 UCF American
61 Akron Mid-American
62 Boise St Mountain West
63 Syracuse Atlantic Coast

 

Bracket Assessment: 20 Days left

The bracket is starting to settle into form this late in the season, with minimal day-to-day fluctuation. We can start to draw conclusions on which teams are already in, which are still in contention, and how many bid-stealing conferences are out there. The field is comprised of 32 automatic bids (from conference tournament champions) and 36 at-large bids.

Tournament Locks

The following 25 teams have already done enough to warrant an at-large bid. Even if each of them were to not win again in the regular season, they would still be selected into the field:

1 Villanova Big East
2 Gonzaga West Coast
3 Kansas Big 12
4 Baylor Big 12
5 Oregon Pacific-12
6 Arizona Pacific-12
7 Kentucky Southeastern
8 Louisville Atlantic Coast
9 Florida Southeastern
10 Duke Atlantic Coast
11 Butler Big East
12 North Carolina Atlantic Coast
13 Florida St Atlantic Coast
14 Purdue Big Ten
15 Cincinnati American
16 UCLA Pacific-12
17 Maryland Big Ten
18 Creighton Big East
19 SMU American
20 Wisconsin Big Ten
21 St Mary’s CA West Coast
22 West Virginia Big 12
23 Notre Dame Atlantic Coast
24 Minnesota Big Ten
25 Wichita St Missouri Valley

1 or 2 wins Away

This next group of 8 will be in the field, barring a monumental collapse down the stretch:

26 Virginia Atlantic Coast
27 MTSU Conference USA
28 VA Commonwealth Atlantic 10
29 Xavier Big East
30 South Carolina Southeastern
31 USC Pacific-12
32 Dayton Atlantic 10
33 Northwestern Big Ten

Remaining contenders

This next group is fighting for the few remaining at-large bids and hoping that the teams in the two groups above them grab all 11 auto-bids in their respective conference tournaments. Some of these teams may be able to sneak in with as few as two more wins, but would be well-served to rack up as many as possible down the stretch to avoid reliance on winning the conference tournament:

34 Illinois St Missouri Valley
35 Oklahoma St Big 12
36 Arkansas Southeastern
37 Iowa St Big 12
38 Nevada Mountain West
39 Virginia Tech Atlantic Coast
40 Valparaiso Horizon
41 Michigan Big Ten
42 California Pacific-12
43 Miami FL Atlantic Coast
44 Michigan St Big Ten
45 Akron Mid-American
46 Marquette Big East
47 Kansas St Big 12
48 Houston American
49 Seton Hall Big East
50 UNC Wilmington Colonial
51 Georgetown Big East
52 Tennessee Southeastern
53 Providence Big East
54 Rhode Island Atlantic 10
55 Boise St Mountain West
56 New Mexico St WAC
57 Vanderbilt Southeastern
58 Pittsburgh Atlantic Coast
59 TCU Big 12
60 Clemson Atlantic Coast
61 Illinois Big Ten
62 Georgia Southeastern

Any team not listed as one of the 62 above will need to win their conference tournament in order to make the field.

Bid-Stealers

Some conferences have teams that would qualify for an at-large if they were to miss out on an auto-bid. This includes the following:

1 Big East
2 West Coast
3 Big 12
4 Pacific-12
5 Southeastern
6 Atlantic Coast
7 Big Ten
8 American
9 Missouri Valley
10 Conference USA
11 Atlantic 10

The top 36 teams above should feel pretty safe even if they don’t win their conference tournament, but outside the top 36 and bid-stealing starts to play a factor. The most important conference tournaments to pay attention to are:

West Coast: Gonzaga or St. Mary’s needs to win this auto-bid to avoid a certain stolen bid.

Conference USA: Middle Tennessee must earn the auto-bid; nobody else in the conference is remotely close to an at-large and the Blue Raiders have all but locked up their at-large.

American: It would be helpful for all if SMU or Cincinnati earns this auto-bid. Houston wouldn’t be bad necessarily, but anyone else would be a disaster for bubble teams.

Missouri Valley: Wichita St is definitely in and Illinois St is very likely in; you want one of these two to get the auto-bid if your team is on the bubble.

Atlantic 10: VCU and Dayton are near-locks for an at-large, so it’d be easiest if one of them earns the A10 auto-bid. Rhode Island is close, so it wouldn’t be that bad if they earned it, but nobody else is in contention for an at-large.

The 6 major conference tournaments are important to keep an eye on, but with so many at-large contenders it is much less likely that a bid is stolen from those.

 

Bracket Matrix update

Yesterday, the NCAA announced that the selection committee would be revealing a preview of the top 16 seeds on February 11th, a month before Selection Sunday. I’ll write a follow-up at that time comparing their top 16 with my own. This preview will be helpful to those of us participating in the Bracket Matrix as it provides a little insight into what the committee views most favorably in their selection process. However, it is not all that helpful to fans since everyone will pretty much already know who are the top 16 teams, give or take a few. These teams are extremely likely to be in the field of 68 a month later as well, so it provides very little insight as to who is on the bubble.

I am hosting my most recent Bracket projections on a static page this year instead of as unique blog posts so they are easier to find. I am aiming to make updates twice weekly until March, at which point I will aim for daily updates. All of my projections are entirely formula-based; I am not interjecting my own opinion or making any adjustments to what the formula projects to be the current field. There will be one exception to this: my final entry for the Bracket Matrix will require some adjustments to account for shortest travel distance for top 16 seeds as well as prevention of intra-conference match-ups on the first weekend. My formula does not currently include this criteria; I hope to automate this for next season but won’t have the time to do so for this year’s tournament.