Mar 13, 2016

Bracket-Picker Checklist (Updated 3/15 5:00pm EST)

I know what you're thinking.....this wasn't supposed to come out until after the brackets were revealed. I can't get my original plan to work, so it will have to wait until next year. I didn't finish my back-up because I watched college basketball all day yesterday. Lastly, this was the absolute worst day of the year to spring forward on clocks and lose an hour (of precious data bracket analysis or sleep, your choice). Anyways, this is a quick article and probably slightly informative, and I can always come back and add more stuff to it.

Anyways, this article should probably be called a cheat sheet because it's eventually going to look like one, even though the presentation may not appear that way.



Upset City

Here in upset city, upsets happen all the time and we keep track of them. So, what is an upset? An upset is counted when the lower-seeded team defeats a higher-seeded team AND the difference between their seeds is 4 or greater. For example, in the first round of games, 6vs11 would be an upset if and only if 11 defeats 6 because the difference between the two seeds is 5 (greater than four). However, 7v10 would never count as an upset regardless if 10 wins because the difference between the two seeds is 3 (less than four). On a technical note, this may be an arbitrary way to look at tournament games because by rule, if a 4 beats a 1, it would not count as an upset, but if a 5 beats a 1, it is counted as an upset. Nonetheless, this is the one law of upset city and we must abide by it. Now, let's look at the results.

The first thing you should notice is that one of my images finally has some color to it (It's about time!). The second thing you should notice is the summaries. I'll start with the totals on the far-right.

Up to this point, my readers should be familiar with the concept of tournament quality. If not, try these links (2016QC-JAN, 2016QC-FEB, 2016QC-MAR, 2016QC-FINAL, This guy just loves to plug his own work.).
  • When a given tournament year features many quality teams, you get tournament stability featured by top-seed advancement (and by rule, a low number of upsets). 2007 and 2009 are prime examples of this phenomenon. 
  • When a given tournament year features fewer quality teams, you get tournament chaos featured by low-seed advancement (and by rule, a high number of upsets). 2006, 2011, and 2014 epitomize this concept, and interestingly enough, upsets in these years continue to happen in later rounds like the Elite 8 and Final Four.
The summaries at the bottom - the averages and the deviations -- may be more important when it comes to picking your bracket. These are the average number of upsets for each round in their respective years, and the value of deviation between the years. On average, the tournament sees
  • five upsets in the round of 64
  • three in the round of 32
  • one in the round of 16
  • one every other year in the Elite 8
  • one every five years in the Final 4
  • zero in the championship game.
The deviations tell the real story. In sane years, the deviations should be deducted from the average, but in crazy years the deviation should be added to the average.
  • Where 4.7 upsets are expected on average in the round of 64, it should be around 3 (4.7 minus 1.7) for sane years. This was close to the case in 2007 when only two upsets happened in the round of 64. When 2.8 upsets are expected on average in the round of 32, it should be around 1.4 (2.8 minus 1.4) for sane years. Again, 2007 fits the mold as only one upset happened in the round of 64. 2009 is definitely a sane year, it just defied the mold. Five upsets happened in the first round (13,12,12,12,and11 won a game), but no more upsets for the rest of the tournament. No other tournament in the list shows that pattern.
  • Where 4.7 upsets are expected on average in the round of 64, it should be around 6.4 (4.7 + 1.7) for crazy years. This was close to the case in 2006, 2011, and 2014 when each of those tournaments produced 6, 5, and 6 upsets (respectively) in the round of 64. Where 2.8 upsets are expected on average in the round of 32, it should be around 4.2 (2.8 + 1.4) for crazy years. Holding true to form, 2006, 2011 and 2014 all had 4 upsets each in the round of 32.
If we can figure out how to define 2016 (sane, average, or crazy), then we should have a pretty good blueprint on how to fill in each round.

Why Kansas should have lost in the B12 tournament

No, this is not a tirade on the officiating (or I'd be talking about the disastrous no-call in the Big East final). So what is the difference between Kansas winning the B12 tournament and Kansas losing the B12 tournament? I think I will let Kansas head coach Bill Self answer this question. In an pre-bracket interview on CBS, Bill Self was asked how does winning a conference tournament affect your momentum going into the Big Dance? In his answer, he addressed the 2008 Title run when they won the regular season and tournament titles, and then his other seasons where the results were mixed (winning the tournament and losing early versus losing the tournament and making a deep run). Then, he said "Losing in the conference tournament can have a way of refocusing a team as well" and I wholeheartedly agree. Here's why:


Sustained success is a very difficult task in college basketball, especially this year given all of the losses by Top 10 and Top 25 teams. These are the winning streaks entering into the tournament of each eventual Champion. If a team entered into the tournament on a loss, the length of the winning streak that was snapped is recorded below the zero. The stand-outs on the list are 2008 Kansas with a 7-game winning streak and 2013 Louisville with a 10-game winning streak. These don't sound like a lot, but for these teams to win a National Championship, they have to add on another 6-game winning streak to their current streak. In other words, these two Champions needed to go on a 13- and 16-game winning streak in order to be Champions. Let's not forget, Kansas needed a Chalmers 3-pointer to go into overtime in a year all four 1-seeds made the Final Four and Louisville trailed most of the way against Michigan in a tight game before pulling it out at the end. I will repeat: Sustained success is a very difficult task in college basketball.

Why does this apply to Kansas? Right now, the rock-chalk Jayhawks from KU are on a massive 14-game winning streak, which means they need a 20-game winning streak to be National Champions. I will not lie: This absolutely scares me. While it is only four more wins than 2012 Louisville had to do, it does concern me greatly. Surprisingly, the KU winning streak isn't even the longest one in the country at the current moment. That honor goes to Stephen F. Austin with an active 20-game win streak, but I don't think anyone has them cutting down the nets. The next longest winning streaks among title contenders are Michigan State on a 9-game winning streak, Oregon on an 8-game winning streak, and North Carolina on a 5-game winning streak.

If this theory is true, here is something to ponder: For 2002 Maryland, 2003 Syracuse, 2005 North Carolina, 2012 Kentucky and 2015 Duke, how would these teams have fared if their winning streaks remained intact heading into the dance. Or as Bill Self suggested, did these losses in conference tournaments re-focus their attention on the big prize?

A New Kind of Streak

The previous discussion was on winning streaks heading into the dance. Is there another type of streak that can identify a possible title contender? Let's try to ask that question a different way: If it takes a six-game winning streak to be a National Champion, have any National Champions went on a 6-game winning streak during their regular season?

To be more practical about it, let's look at every national champion since 2000 and see if they completed a 6-game winning streak against power conference teams (after all, the tournament is pretty much a 6-game winning streak against that level of competition).

In the chart above, the P/F row (Pass/Fail) shows the longest winning streak against power conference teams. 13 out 16 champions passed this test. Two of the champions have an asterisk beside their streak, and this denotes streaks that were interrupted by non-power conference opponents. For example, the 2006 Florida Gators went on a 14-game winning streak from Nov to Jan, in which 8 of these opponents, though not consecutively, were power conference teams. Likewise, the 2010 Duke Blue Devils went on 7-game win streak during ACC play, but for some reason, they scheduled major conference Tulsa as the 7th game followed by their 7th ACC victory (8 games total in the streak).

What about those that failed? Well, I tried an alternate test: Did they beat 5 straight power-conference teams, but this streak must include the conference tournament (to account for tournament-style play)? One of the three teams passed this test (2000 Michigan State). The two teams that failed this test were 2004 Connecticut and 2014 Connecticut (which has made me come to the realization that when things are crazy, just go with Connecticut). The three teams in the Alt row that are marked with a "+" indicate teams that also met this criteria, even though they passed the first test.

The real question: How many teams in the current field pass this test:
1-seeds: KU+, UNC+, UVA+, ORE+
2-seed: MIST+, XAV*, NOVA+
3-seed: WVU+, TXAM+, UTAH+
4-seed: (UK), CAL
5-seed: IND+, (PUR)
6-seed: ARI
7-seed: IOWA, WISC

Notation
NAME = Pass 1st Test
(NAME) = Fail 1st, Pass 2nd
NAME+ = Pass Both
*=Pass 1st Test, but streak was non-consecutive games

Tournament Performance

The biggest decision that a bracket picker has to make is predicting the eventual champion. Followers of Bracket Science will remember that Pete used a rule-based system to pick the Final Four and Champion, and did so with amazing accuracy. In the previous section, I approached the decision by asking the obvious question: Can this team go on a 6-game winning streak? After looking through my notes, I should have asked this question: Can this team win at least one conference tournament game? It kind of makes intuitive sense. How can you expect to win six straight games against the best of the best when you can't even win one game against a team which you should already be very familiar? Every national champion since 2000 has not lost the first game of its conference tournament. Many have lost the second game in their conference tournament (02 Maryland, 03 Syracuse, 05 North Carolina, 09 North Carolina, and 15 Duke), two have lost in the championship game (12 Kentucky and 14 Connecticut), and the rest have won their conference tournament. None have lost their opening game! So, which teams in this year's NCAA field have lost the opening game in their conference tournament?

4-seeds: Iowa State
5-seeds: Indiana
6-seeds: Texas, Wisconsin
7-seeds: Iowa
8-seeds: Texas Tech

Even though no team lower than an 8-seed has ever won the NCAA title, I'll list them as well.

9-seeds: Butler
10-seeds: Syracuse
11-seeds: Tulsa, Vanderbilt

No comments:

Post a Comment