Quality Curve Analysis (February Edition)

The Idea Behind Quality Curves

As stated in the January Edition, the theory states that the quality of the teams in the tournament can predict the quality of the tournament results. In layman's terms, higher quality teams in the higher seeds will result in a stable, predictable tournament (much like 2007) and lower quality teams in the higher seeds will result in an insane, unpredictable tournament (much like 2014).

So let's start things off with me giving myself a huge pat on the back, and how do I do that? By seeing how my prediction in the previous post turned out. I said:

My take: At this point in the season, all but 2 conferences have fully begun conference play (Atlantic Sun and Ivy League). As teams began conference play, I expect the 2016 QC to bend downward as power conference teams face higher quality competition on a night-in, night-out basis. I also expect the 11- and 12- seed teams to slightly increase as these seeds may be composed of mid-major and minor conference teams whose conference play features teams that power conferences play against in their non-conference schedule. Thus, the better teams in these mid-major and minor conferences should improve their ratings as they separate themselves from the weaker teams in their conference. When I look at the 2016 QC again in February, we'll see how if this prediction comes true.

Did this prediction come true or was I way off base? Let's combine the two QCs onto one chart and find out.

Money in the bank! But to be completely objective about it, this was like predicting if the sun is going to rise in the morning. Of course it was going to rise, it was only a matter if we were going to see it. The reasoning I gave as to why I thought the QC was going to bend in that particular matter is actually self-explanatory if you think about it. Anyone could have made that prediction.

• What I actually find most interesting is the inflection point, or the point on which the two curves seem to rotate. The inflection point of the Jan QC and the Feb QC appears to be the 4-seed. If I had predicted which seed was going to be the inflection point, that would have been something to brag about. 
• The second aspect I find most interesting are the deviations between the two curves. 5s, 6s, 7s, 11s, and 12s seem to have made the largest moves.

QUICK ANALYSIS OF FEB QC

1. The obvious point of emphasis on the 2016 QC is how the 1s and 2s compare to the other QCs. The 1s are already below the 2014 Insanity Curve and the 2s are below the 2007 Chalk Curve. What's even worse is that these are the four best and four second-best at 1 and 2, respectively. If the Selection Committee doesn't hand out seeds in this exact order (Spoiler: They never do!), then the 2016 QC will get pushed down further at these two spots. It doesn't look like you want to be a top-seeded team in the tourney this year (and it hasn't been pretty to be top-ranked during the regular season).
2. Seeds 3 thru 7 are abnormally high compared to both the Chalk and Insanity curves. The 2016 QC doesn't seem to take a sharp dip until the 8 seeds (which is more visible on the Jan-Feb Comparison). This could be a sign that upsets may not happen until the Sweet Sixteen rounds when 1s play against 4s or 5s and 2s play against 3s, 6s, or 7s. Of course, we'll have to wait and see how the teams actually get seeded before we make any predictions like that.

Although you can't see the actually data values, I will provide the decline values between seeds like I did in the January Edition.

• From 1- to 2- seeds, the decline is .015.
• From 2- to 5- seeds, the decline is .009 for each seed group.
• From 5- to 7- seeds, the decline is .014 for 5 to 6 and 0.012 for 6 to 7.
• From 7- to 8- seeds, the decline is .022.
• From 8- to 9- seeds, the decline is .008.
• From 9- to 12- seeds, the declines range from .011 to .013.

This is a much more sporadic differential than those in the January Edition. In the first QC Analysis, the narrow differentials occurred from the 6- to 9-seed groups, yet in this one, the narrow differentials occur in the 2- to 5-seed groups and in the 8- to 9-seed group. I'll definitely pay close attention to these differentials throughout this month of games to see if it reveals anything useful for predicting the bracket.

A Bracketology Curve

Since our QC follows a top-down ordering of teams, I thought it would be fun to make a QC for an actual bracket, called a Seed Curve (SC). (NOTE: I have edited previous articles to emphasize the distinction between QCs and SCs). Since teams rarely get seeded according to efficiency ratings like KenPom Ratings, SCs tend to fluctuate up-and-down rather than fall in a straight line like a QC. To do this, I will build a curve based on where teams are seeded according to the latest Bracketology predictions (for a definition of Bracketology).
NOTE: I am partial to Bracketology that predicts the tournament as it would be on Selection Sunday, not the Bracketology that predicts the tournament if the season ended today. Since the QC analysis performed above doesn't take into account games not yet played, I am forced to use a type of the latter. Being so, I chose ESPN's Bracketology since it accounts for all games up to Jan 31, just like the QC Analysis.

As you can see, the SC looks a lot different than the top-down QC above. Let's take a closer look and see what we could be dealing with in March.

• Below-2014 1-seeds and middle of the road 2-seeds. Contrast that with much stronger 3-, 5-, and 6-seeds. From the 7- to 12- seeds, the SC seems to match the 2014 SC except noticebly stronger 9s and 12s.
• In comparison to the QC which takes the Top 50 ranked KenPom teams and seeds them accordingly, the SC featured seven teams outside of the Top 50 KenPom. 8-seed Colorado checks in as the KenPom #51 team, which is the equivalent to a 13-seed. 10-seed Washington checks in as the KenPom #76 team, the lowest-ranked KenPom team to be projected into the tournament, and it corresponds to a 19-seed, suggesting they shouldn't be in at all. The other five teams are LSU (10-seed, #59 KenPom = 15-seed), Clemson (11-seed, #57 KenPom = 14-seed), UCLA (12-seed, #62 KenPom = 15-seed), Monmouth (12-seed, #58 KenPom = 14-seed), and San Diego State (12-seed, #60 KenPom = 15-seed).
• A instant glance at the SC shows us that 2014 is an very close match to 2016. Looking back at the SCs here, I would say that 2006 is the next best approximation for 2016. If you remember back to the 2006 Tournament, the only Final Four team with a 1 in its seed-number was 11-seed George Mason. With 2014 and 2006 being two of the upset-ridden tournaments, it seems like we can expect a similar situation in 2016.

This concludes the February Edition of Quality Curve Analysis.

1. I will do this again on Feb 29 to account for all games through Sun Feb 28 (the March Edition). If I feel like the QC is taking shape or making drastic changes in certain seeds, I may do an interim update. Then, I will do the actual 2016 SC on Mon Mar 14 after the brackets are revealed on Sun Mar 13 (and all regular season games have completed).
2. For a detailed explanation of QCs and SCs, see this article.