M Hockey: Midseason update and a look ahead

The hockey team is now roughly halfway through the 2011-12 season.  I thought this would be a good time for a recap of the first half, along with a (very) early look at the pairwise rankings heading toward NCAA Tournament selection time.   By the way, thanks to MGoBlog for already having a "way too early to be doing this" tag. 

First half recap

Michigan got off to a strong start, at one point holding down the #1 ranking in the polls with a 4-0 record following easy wins over Niagara, Bentley (twice), and St. Lawrence.  A trip to Northern, resulting in a loss and a shootout win, put a quick end to that.  After winning three of four against Ferris and Western, Michigan went on a massive losing streak, with one point in the next seven games, before a 1-0 OT win in Alaska.  After getting four of a possible six points two weeks ago against MSU, Michigan now sits 8th in the CCHA with 19 points in 14 games, and has a lot of work to do to get back into the NCAA tournament mix.  The defense has been mediocre at best, and Jon Merrill's absence has really hurt.  All signs are pointing to Merrill being reinstated for next weekend's matchup against the Lakers, but we'll have to wait and see.

Current Pairwise Rankings

For the uninitiated, the pairwise rankings compare every team to every other team in four categories: RPI; record against “Teams Under Consideration (TUCs)”, which means teams with an RPI over .500, as long as both teams have played at least 10 TUCs, not counting each other.; record against common opponents, and head-to-head.  Winning one of the first three categories gets you one point, and you get one point for each head-to-head win over your opponent.  Any ties are broken by the RPI ranking.  The pairwise rankings have accurately predicted the NCAA field for something like ten years, so they are a very useful and important tool. 

As an example of how the pairwise rankings work, let’s compare Michigan to Miami right now. 

• Michigan has an RPI of 0.5374 to Miami’s 0.5264.  Point Michigan.
• Michigan is 5-7-2 against TUCs other than Miami.  Miami is 7-9-0, for a slightly better winning percentage.  Point Miami.
• Michigan is 3-4-1 against common opponents.  Miami is 3-5-0 against those opponents.  By a very slim margin, point Michigan.
• Miami has one head-to-head win over Michigan.  Point Miami. 

This puts the comparison at 2-2.  Michigan wins the overall comparison by virtue of having the higher RPI score.

Every team is compared to every other team in the top half (essentially) of the RPI standings, and the teams are ranked based on the number of comparisons won.  Right now, Michigan sits 17^thin those rankings.  I think that based on the really awful stretch Michigan had in November, many would have felt that they would be further down the list, but the CCHA teams are ranked very high in the Pairwise right now based on their incredibly high RPIs.  In fact, 9 CCHA teams are in the top 19 in the Pairwise as of this writing.  Right now, with very few teams outside of the CCHA having ten games against TUCs, this is a heavily RPI-driven ranking.  That will change as the season goes on, and in a bad way for Michigan unless they pick up some big wins down the stretch.

Why is the CCHA doing so well in the pairwise rankings? 

It turns out that nonconference games have a huge impact on the Pairwise for a couple reasons.  First, having a conference do well overall in its nonconference slate has a big impact on the RPI, because it improves every conference team’s strength of schedule.  The CCHA is 40-12-5 out of conference so far this year, for a winning percentage of 0.746.  Even terrible Bowling Green, who is 1-11-2 in conference, went 5-0-1 in nonconference play, which helps everyone else in the conference. Secondly, when comparing teams from different conferences, the common opponents part of the Pairwise is usually based on a very small number of games.  For example, if we compare Michigan to a Hockey East team such as Merrimack, the common opponents are Boston College, Northeastern, St. Lawrence, and Union (Merrimack won’t play St. Lawrence until Friday, and neither has played BC, but both will). Right now, Michigan is 1-2 against those teams, while Merrimack is 2-1-1 against them (mostly against Northeastern).  Had Michigan beaten either Union or Northeastern, it would have made a huge difference.  As it stands, beating BC in the GLI would be a huge help in comparisons against all the Hockey East teams, since they will mostly have bad records against BC.

Second half preview

We start with the GLI, and a matchup with #3 Boston College, followed by either MSU or a surprisingly competent Michigan Tech team.  The next six series are all against CCHA teams that are in the top 20 in the pairwise, including four (Ohio, Notre Dame, Northern Michigan, and MSU) that are currently in the top 10. The season ends with a two-game set at Bowling Green. 

This is both good and bad.  Obviously, the schedule down the stretch is pretty brutal, with the Ohio and ND series on the road, and the MSU series as an away/neutral split.  The good news is that this gives Michigan a huge opportunity to work its way back into the mix for an NCAA spot and a higher conference standing. 

How does Michigan get back to the NCAA tournament?

Well, winning the CCHA tournament would be nice, but I’m not holding my breath.  Michigan is going to have a fairly strong RPI for its record no matter how the season ends up; what could hold Michigan back is their sub-.500 record against TUCs, and their losses to Union and Northeastern, which (along with some other losses to CCHA teams) are hurting them in the common opponents comparison in the pairwise rankings.  With only the GLI remaining outside the conference, Michigan is going to have to rely largely on improving its record against TUCs.  The good news is that there are still plenty of those left on the schedule.  With at least 13 regular-season games against TUCs, plus a couple in the playoffs, Michigan should have 15-17 more games against TUCs.  Right now, Michigan has a record of 5-8-3 against TUCs.  That winning percentage is not going to be good enough, especially given Michigan’s poor record against out-of-conference opponents.  They’re going to have to get that TUC record up to at least .500, and probably higher. 

To start with, Michigan needs at least a split in the GLI.   A loss to BC won’t hurt too badly, but a win would be huge.  After that, assuming a sweep against BG to end the season, Michigan is going to have to go through that 12-game CCHA stretch following the GLI with at least a .500 record, and it might be necessary to go as much as 8-4.  While the RPI numbers are going to be good for Michigan, the TUC record is going to hurt them down the road if they don’t win more than half of their games the rest of the way.

A lot of things can change between now and March, aside from Michigan's performance.  Every game over the next two months is critical, and it's going to be an exciting ride as Michigan tries to extend its consecutive tournament appearance streak.