2019 Quality Curve Analysis - March Edition

Well, the calendar has rolled over to another month, and at PPB, it means it is time for another installment of the Quality Curve analysis, so let's see what's changed in the last month.

The Evolution of the Quality Curve

Let's take a look at the current KenPom QC, and see how it has transformed in one month's time.

I tried to keep the same color patterns as the February QC (yellow for Jan and light blue for Feb), and I added pink for March to give it an Easter theme. Anyways, the curve is still staying strong in all areas except for the #24-#27 ranked teams. These would be potential 7-seeds, but it is most likely these teams do not end up as 7-seeds. For the most part, the QC is holding to form, meaning no irregularities and no drastic shifts or rotations. The curve did take a noticeable movement downwards in the middle of February, but has recovered to its early-February heights. This movement can be noticed in the next chart.

This chart is the same chart above, except the maximum and minimum curves for the month of February have been added in black. Since both the Feb (light blue) and Mar (pink) curves are 96-98% bounded by the black max- and min-curves, it tells us that our current reading (all games up to Feb 28) is an accurate reading across the whole month instead of being only an accurate reading for the specific point in time (the end of the month). Although it isn't shown in the chart, the only areas of the curve where the February max-min range is narrower than the March max-min range is the #18-#28 spots in the QC. Why is this interesting? The area of weakness identified earlier (#24-#27 spots) is contained in this range. In other words, this area of the curve didn't move up in January when the rest of the curve did, nor did it move much in any direction during February. This area is weak and remained weak. As usual, let's take this same approach with a different data set.

Above is the current Sagarin QC. While we don't see much separation between the three curves in the Sagarin QC like we saw in the KenPom QC, we still see the March QC showing the same consistencies with the February QC rather than the January QC (this was also the case for the KenPom QC). It simply means that our current QC is an accurate reflection of tournament quality across the entire month, not just this specific moment in time. The separation at the front of the curve (#1-#6 spots) is indicative of stronger top seeds. When top seeds are playing at their peak near the beginning of March, it is a good sign of no upsets from 15- and 16-seeds (yes, we have to take 16-seeds into consideration thanks to last year). The separation at the back of the curve (#27-#50 spots) is indicative of lower-seeded teams gaining momentum and possibly looking to upset 3- thru 6-seeds. Remember, nothing is definitive until our seed-curve is produced in two weeks. Now, let's look at the Sagarin QC with the max-min curves.

This chart confirms the same points above: Separation at the front of the curve and the back of the curve. Although, the points of max-min widening are interesting. There is a narrow spot of widening at the #21-#23 spots, which is in the same area of widening as the KenPom QC. The other area of widening is at the #35-#38 spots. This would area of the curve would represent the 9- and 10-seed range of teams, but it is interesting to note that the March QC (pink curve) is at the mid-point of the range from the #35-#37 spots and plummets toward the minimum curve at the #38-spot. If these teams are going to spring upsets in March, it will come as a surprise since the QC indicates they aren't playing their most efficient at the current moment.

The Message of the Quality Curve

If the charts in the last section told us that the top teams were playing at or near peak-efficiency in March, this chart tells us what they are likely to do.

This chart is a year-by-year comparison of the 2017, 2018 and 2019 KenPom QCs in March. Here is what I see.

1. The top seven teams are more efficient than 2017 and the #8-#11 teams are just below the 2017 curve. What does that mean? In 2017, all pod leaders (1- thru 4-seeds) won their opening round game. If the parallels in the two QCs hold to form, 1- and 2-seeds (assuming the top seven in the QC get either a 1- or a 2-seed) and a few of the 3-seeds (assuming the #8-#11 in the QC get 3-seeds) should be safe from R64 upsets.
2. From the #12-#26 spots, these teams are either average (splitting the difference between 2017 and 2018) or 2018-compatible. How would I interpret this? In 2017, one 1-seed and two 2-seeds fell in the R32 to an 8-seed (an under-seeded 8-seed I might add) and two 7-seeds, respectively. If 1- and 2-seeds are facing historically strong 7- and 8-seeds in the R32, I would probably expect one or two of the top seeds to fall. In 2019, the 7- and 8-seeds do not look as strong as they were in 2018, so logically, I would assume more of the 1- and 2-seeds to survive into the S16.
3. The teams #28-#45 are also historically strong, which represents the 8- to 11-seed range of the curve. How would I interpret this? Both 2017 and 2018 saw a lower-seeded team (7-seed in 2017 and 11-seed in 2018) make the F4. Considering every year since 2011 except for 2012 has produced a F4 Cinderella team from this range of the curve, I would think the historically strong quality of teams in this range increases the probability that it happens in 2019 again. With the strength at the very front of the curve where the 1-seeds can be found, I doubt the run will come from an 8- or 9-seed. If I had to make a guess before the bracket comes out, I would say either a front-of-the-curve team gets under-seeded as a 7-seed (i.e. - VT) or one of the historically strong teams in this area of the curve (#28-#45 spots) gets seeded as a 10- or 11-seed (i.e. - WOFF) or maybe even as a 12-seed play-in team (i.e. - CLEM). Considering the relative weakness of the 3- thru 5-seed range, I would bet my money on the latter.

As usual, we will sort all of this out in two weeks when the bracket is revealed and we can look at the final QC as well as the 2019 SC. Now, let's look at the Sagarin QC year-to-year comparison.

This chart shows us the same message as the KenPom QC year-to-year comparison, albeit with some slight variations. Let's dig deeper into those variations.

1. The #1-#5 teams are either ridiculously strong or in-line with the 2017 curve. This is a good sign for 2019's 1-seeds for the R64. (we know what happen to a 1-seed in 2018). The #6-#8 teams are in-line with both curves, which is definitely a good thing for 2-seeds in the R64, as all four won in the R64 in both years. 
2. The #9-#16 teams matches up with the 2018 curve, so we need to pay attention to where these teams are seeded. They could be potential upset-victims if seeded in-line (3- or 4-seeds) due to their historical weakness.
3. From #17-#27 (5- thru 7-seeds), these teams seem to split the difference with two curves. I would again assume this to mean teams in this range are potential feasts for 1- and 2-seeds if seeded in-line with their ratings.
4. From #28-#37 (8- and 9-seeds), we see the historical strength that we saw in the KenPom year-by-year comparison. This seems to point towards the "under-seeded 10- thru 12-seed" theory as the likely candidate for a Cinderella F4 run in 2019.
5. The back of the curve (#41-#50) collapses to historical weakness, but it is very likely that teams in this range do not receive invites to the dance, and if they're not in the dance, they are of no concern to us.

For predictive purposes, I like what the 2019 QCs are showing us. We have two weeks until we get the final look at it, but until then, things are looking very promising for us to achieve our goal of a perfectly predicted bracket. As always, thanks for reading my work, and I'm planning to post another article on March 11, as well as the usual works during Bracket Crunch Week. I hope to see you then.