We're just past the halfway point of the season. Here's some stuff about stuff.
Fans of Northern schools think that Mid-Atlantic and Southern schools are always overrated. Fans of Mid-Atlantic and Southern schools don't know where Northern schools are (despite getting the weather of Northern schools this season). And Johns Hopkins fans think that everyone is overrated because (1) they haven't won 5,735,185 titles in their history, and (2) nobody has come to their senses and made a simple "Johns Hopkins is Number One!" poll that includes only Johns Hopkins, ranked first (of course).
The concept of overrated and underrated teams is weird. There are lots of factors that go into that determination, and I don't have the time or desire to really iron out hard answers to that issue. Instead, I tend to gravitate toward addressing that issue in its most simplistic form: Looking at the human polls, which teams are people voting higher or lower than what cold, calculating computers think? It's not a perfect resolution to the issue, but it's pretty good for our purposes.
In that analysis, four teams stick out as potentially holding a position in the human polls that doesn't comport with what the machines -- our future overlords -- think. The below aggregation considers every team slotted in the top 20 of at least one set of rankings. Some brief thoughts follow the table.
LAXPOWER: These are the LaxPower ratings. These ratings are based on solid math, importantly considering margin of victory.
RPI: This is stupid person math. I include these rankings only because the NCAA is full of stupid people and they still use this stupid person math as a major piece of their stupid tournament selection criteria. I've included these rankings because I am, apparently, as stupid as everyone and everything else.
EFFICIENCY: This is just a ranking of a team's Pythagorean win expectation. These are similar to Ken Pomeroy's rankings, but slightly different.
AVG.: Average of the "math" rankings.
COACHES/MEDIA: These are the human polls, as voted on by humans. These humans have different jobs, though: One set judges humans while clad in university-issued apparel; the other set of humans judges others simply to sell ink and paper. These polls are from March 24, 2014.
AVG.: This is the average the Earth-human polls.
DIFF.: The difference between the "math" polls and the human polls. A negative value means that "math" is rating a team higher than the things that are carbon-based; the inverse means that the things made out of 75% water are rating a team higher than the things not made out of any water.
AVG.: Average of the math and human polls. This is how the table is ordered.
Some brief thoughts on all of this:
- Pennsylvania: Every season it seems like the Quakers are given higher value in the computer-type rankings than in the human-type polls. Pennsylvania is a good team, the most undervalued team in the human rankings, and are better -- in totem -- than their record. The Quakers started the season ranked 15th in the media poll, fell back, surged forward, and are now back to where they started. Yet, the team's performance on the field -- despite the losses to Duke, Princeton, and Cornell -- is greater than the team's human poll ranking. The Quakers are merely the victim of the sensibilities that drive human poll balloting. There's still lots of value in Penn, even if that value hasn't been fully realized in the team's record.
- Bryant: The Bulldogs are on the opposite end of the spectrum from Pennsylvania. Bryant is arguably the most overvalued team in the polls, holding computer-type ratings that are significantly lower than their human poll positions. The team's record is gaudy -- 7-2 with five consecutive victories -- and there are a handful of good wins in Bryant's bucket, but the Bulldogs aren't bulletproof: The team is 3-2 against opponents expected to win at least half of their games and have four wins against teams expected to win fewer than half of their games. Does that sound -- without question -- like a team that's top 12 quality? The Bulldogs have that kind of potential, but they haven't created a resume exactly in line with their human poll rankings.
- Cornell: Ranked slightly above their head, but the Red are undefeated. The team's positioning in reasonable, but it's important to note that there is some concern about the Red -- based on ranking difference -- in the computer-type polls. (The Red's adjusted defensive efficiency value ranks 22nd nationally. That isn't a death blow for Cornell, but it is a red flag for a team approaching the top spot in the human polls.)
- Harvard: The lack of love for the Crimson is weird. The Crimson have been on my "almost in" list for a few weeks, but Harvard continues to receive very few votes in both human polls. This is a team ranked 19th in adjusted offensive efficiency, 14th in adjusted defensive efficiency, has significantly improved in its biggest area of concern from 2013 (controlling turnovers), and is sticking it to suckers (even if the Crimson have struggled to topple their best competition). Harvard has the smell of a team that could make a push in slate as the season continues to develop. This is as strong as the Crimson have been in a while.
- General Thoughts: There really isn't a lot to complain about with respect to teams being over- and underrated. 72 percent of the table has a ranking difference under three positions. That's . . . kind of amazing. And if you throw out teams with an average poll position of 21 or higher, 84 percent of the table has a ranking difference under three positions. If you've trusted the human polls this season, both measures have -- on the whole -- delivered (at least through the last set of human polls). The RPI is a totally difference story: Kill it with fire.