Tuesday, December 10, 2013

Re-Education Through Labor

A few weeks ago, we mentioned in passing while discussing the Engineers' less-than-expected luster on their non-conference schedule that the PairWise Ranking calculations had changed fairly drastically.

If you've been following WaP over the last couple of seasons - more accurately, in two of the last three years - you should at least be somewhat familiar with the PWR. In the 2010-11 season, when the Engineers reached the NCAA tournament, we broke down week-to-week from January through to the tournament what the rankings looked like, and we did the same last February when it appeared that the Tute might be destined to overcome a terrible start and make the tournament (they didn't). Engineer Bracketology is a "micro-bracketology" effort that focuses on the Engineers and what they wanted to see happen outside of its own games in order to bolster either its chances of making the tournament or its potential seed.

This year, however, there's been a radical change to the components of the PWR, and the biggest one has to do with what was always a major topic in these micro-bracketologies: the TUC cliff.

Prior to this season, there was a distinction known as Teams Under Consideration, which ostensibly separated good teams from bad teams. Depending on the year, a TUC was either a team in the Top 25 of the Ratings Percentage Index (RPI) or a team with an RPI of .5000 or higher. That category has now essentially been abolished.

What does that mean? It means that instead of the PWR consisting only of TUCs, it now consists of comparisons between all 59 Division I programs.

The other three elements of PWR - RPI, common opponents (COp), and head-to-head (H2H) - are the remaining components of each comparison, of which there are now a total of 1,711 as each team is compared to every other team in the nation. This won't cause a great deal more number-crunching, realistically, since A) there's still only 16 teams in the tournament and B) the teams that will be on the bubble for an at-large bid are almost certainly going to be winning comparisons with all the teams south of at least 30th or so, barring extreme circumstances.

RPI will now more heavily weight wins on the road (by a factor of 1.2) than wins at home (0.8). Neutral ice games are weighted normally. The inverse is rue for losses at home, weighted 1.2, and losses on the road are weighted 0.8. Ties are treated as half a win, and half a loss.

RPI is also now being injected with a new bonus for quality wins. For victories over the top team in RPI, a bonus of .0500 is added. For the #2 RPI, a bonus of .0475, and so on through #20, which provides a bonus of .0025. The standard weighting for road/home/neutral ice is then applied to the bonus, and ties are again half a win, half a loss, so a quality tie produces half of the bonus it otherwise would.

So in other words, we're still going to be watching a specific set of teams - the TUC cliff has essentially been replaced by the RPI cliff, but its impact will not be nearly as wild as the TUC cliff.

There's also the numbers game, given the new lay of the land in college hockey. Last year, there were only five conferences, and thus only five automatic bids against 11 at-large bids. Now, with the demise of the CCHA and the rise of the Big Ten and the NCHC, there are six automatic bids. That changes the scope of what the PWR positioning will mean. Much of this assumes that the Atlantic Hockey champion will not be in the Top 15.

1-10: In the tournament, no matter what.
11: Almost certainly in the tournament. There would have to be upset champions in every conference for this team to be out.
12: Also likely in the tournament. Four upset champions put this team out.
13: The good side of the bubble. Good odds of being in the tournament, but three upset champions put them out.
14: Firmly on top of the bubble. Two upset champions put them out.
15: The bad side of the bubble. Just one upset champion would put this team out.
16: Out of the tournament, unless the Atlantic Hockey champion is in the Top 15 AND there are no upset champions in other conferences.
17-59: Out of the tournament without an automatic bid.

No, we're still not getting fully into PWR just yet. Why? The college hockey season unfolds over the course of five and a half months from the beginning of October through the middle of March, and while the PWR gets more important as time goes on, it doesn't really become an indicator of anything resembling a predictor of national tournament strengths until February, although it does become interesting to look at starting in a few weeks.

Here's my quickie chart for how to characterize the PWR during each month of the college hockey year:

October: Worthless.
November: Farcical.
December: Quirky.
January: Interesting.
February: Important.
March: Essential.

So we're still here sitting on quirky. There are some clearly good teams being indicated, now in the brave new PWR world, there are some clearly bad teams being indicated, but the murky middle tells us almost nothing at this point. If, following the Engineers' trip to Minnesota, they're in the PWR picture for an at-large bid, we'll fire up Engineer Bracketlogy, but not before.

However, if you want a look at what the Engineers have in terms of quality thus far, here's what we've got: two home quality ties against Quinnipiac (#8) and Cornell (#10), and a home quality win against New Hampshire (#17). Three losses (one at home, two on the road) to Boston College (#7) and Union (#9) are missed opportunities. There are games coming up in Minnesota that are likely to be possible quality opportunities against Ferris State (#4) and either Minnesota (#2) or Colgate (nearly a quality opponent at #24).

No comments:

Post a Comment