clock menu more-arrow no yes mobile

Filed under:

Hockey Bubble Talk: March 2-4 Edition

Hockey is off this week thanks to a split against Notre Dame, clinching a top 5 spot in the CCHA and a bye in the first round of the CCHA playoffs. Unfortunately, since we ended up in 5th, we will be on the road for the quarterfinals against Miami next week. Thanks to the Pairwise Rankings, however, we can make a somewhat more educated guess as to where things stand than in basketball. A reminder of how they work (which is slightly different from two years ago; I'm not sure whether the changes were implemented last year or this year, since we had no chance of making it last year):

  • Shootouts don't exist as far as the Pairwise Rankings are concerned. Any game that ended in a shootout (CCHA regular season games or any in-season tournaments which use the shootout to decide advancement rather than playing open-ended OT) is considered a tie.
  • Teams with an RPI of .500 or above are "under consideration". (This is a change; two years ago the cutoff was top 25. At the moment there are 32 teams under consideration, or TUCs.)
  • Every TUC is compared against each other TUC as follows:
    RPI: Higher RPI gets one point. RPI is also the tiebreaker in the event that the total comparison is even.
    TUC: Better record against fellow TUCs (not including head-to-head matchups, those are handled separately) gets one point, as long as both teams have played 10 such games. If one or both fall short of that threshold, neither team gets a point.
    Common: Records are compared against common opponents. In previous years, the aggregate winning percentage was used even if (as was common) the number of games against various opponents was not equal; this could be a major disadvantage if you played, say, five games against Michigan and one against Michigan Tech (usually a bottom-feeder in the WCHA, although they're semi-competent this year) and someone else played five against Tech and one against Michigan. This year the common-opponent point is decided differently: the winning percentage against each opponent is computed separately and the winning percentages are added together.
    H2H: Each win in head-to-head play gets you one point.
    The team with the most points from this comparison wins the comparison; if they have an equal number, RPI is the tiebreaker.
  • Teams are ranked by the number of comparisons won. Ties are broken by RPI.

Right now we sit in 11th place (tied with Northern Michigan and Denver on comparisons won, but with the highest RPI of that trio). 16 teams get in; this usually includes one or two auto-bids who are outside the top 16 in the Pairwise, so the cut line is usually at top 14 or 15. A detailed breakdown of what we are rooting for this week follows after the jump.

As a rough estimate, a .500 team facing an opponent equal to their average SOS will move their RPI about .0040 in either direction; a team above .500 will gain less or lose more proportionally. Playing an opponent about .250 above your average SOS is worth enough of a boost that a .500 team would come out even with a loss. This is mostly for estimating how much RPI movement is possible.


Very Unlikely to Win

Boston College (.0240 back in the RPI and lost head-to-head)
Michigan (trail in every category and -2 head-to-head)
Ferris (same situation as Michigan)
Massachusetts-Lowell (clinched common opponents by beating BC once and has a large TUC lead)

Comfortable Wins

Western Michigan (.0169 RPI lead and +2 head-to-head)
Ohio State (.0193 RPI lead, +2 head-to-head)
Colorado College (huge RPI lead and guaranteed common opponents)
Harvard (huge RPI lead, no common opponents)
Quinnipiac (huge RPI lead and guaranteed common opponents)
Wisconsin (huge RPI lead, common opponents appears to be clinched barring an insanely specific set of outcomes)
St. Cloud State (huge RPI lead, guaranteed common opponents)
Northeastern (huge RPI and TUC lead)
Air Force (RPI + head-to-head)
Nebraska-Omaha (RPI + common opponents)
Bemidji State (RPI + TUC)

Up for Grabs

Minnesota-Duluth (L)
RPI gap: -.0096
TUC: 14-15 (MSU) vs. 12-6 (UMD)
Common: Notre Dame (both 1-1), Michigan Tech (1-0 vs. 2-1-1), Western Michigan (both 2-0), Minnesota (1-0-1 vs. 0-2); we can't lose the lead here (even with losing to ND and Western at the Joe and UMD beating Minnesota and Michigan Tech in the WCHA tournament).

Amazingly, even though they're #1 in the Pairwise, a relatively small shift in the RPI would give us the comparison.

Boston University (L)
RPI gap: -.0061
TUC: 14-15 vs. 15.5-9.5
Common: Boston College (0-1 vs. 2-2), Notre Dame (1-1 vs. 0-1); currently tied, but a win over ND at the Joe would win it for us (conversely, a loss would make winning common opponents impossible). If we don't play ND, a BU-BC meeting could have the same effect (but because they've played more games against BC than we have against ND, an extra game against ND would take precedence).

RPI is within swing range; that would be enough as long as BU does not face BC and we don't face ND. If we play ND, we have to win; if we don't play ND and BU plays BC, BU has to lose. (In either case, we have to get RPI.)

Minnesota (L)
RPI gap: -.0033
TUC: 12.5-14.5 vs. 14-8
Common: Michigan Tech (1-0 vs. 1-1), Notre Dame (1-1 vs. 0-1); we've clinched this.
Head-to-head: +1

RPI will decide this one, as the RPI gap will be far easier to overcome than the TUC gap.

Miami (W)
RPI gap: +.0013
TUC: 14-13 vs. 13.5-13.5
Common: Our entire conference schedule; we have a narrow lead here (+0.175).
Head-to-head: -2

Get swept and we lose this no matter what. Lose 2-1 and we actually still have a tiny chance, should Miami lose both games at the Joe (they'll be 2-3 vs. our 1-2 in those games, so RPI should be fairly close; we'll still own TUC and common). Win the series and we probably still have to win one at the Joe (or at least tie the third-place game) to avoid dropping the TUC and common points (although if the games at the Joe are against teams that swept us, they can't hurt our common opponents record).

Union (L)
RPI gap: +.0018
TUC: 14-15 vs. 8.5-6.5
Common: Western Michigan (2-0 vs. 0-0-2), Michigan (1-3-1 vs. 1-0). No way to win this.

Only chance is to make up the gap in TUC.

Maine (L)
RPI gap: +.0051
TUC: 14-15 vs. 11.5-11.5
Common: Boston College (0-1 vs. 2-1). No way to win this.

Need to make up the TUC gap here as well; this is a little less daunting than the gap against Union.

Northern Michigan (W)
RPI gap: +.0070
TUC: 11.5-13.5 vs. 12-12
Common: +0.550
Head-to-head: +1

Keeping RPI is enough if we don't play them. If we do and we lose, we need to maintain both RPI and common, which isn't likely.

Denver (L)
RPI gap: +.0094
TUC: 14-15 vs. 14.5-10.5
Common: Air Force (both 1-0), Boston College (0-1 vs. 1-0), Minnesota (1-0-1 vs. 2-0), Michigan Tech (1-0 vs. 0-1-1), Miami (0-2 vs. 0-1)

A sweep of Miami would be very helpful here; if Denver does not play Minnesota or Tech, that would make common opponents a tie (and give us the win on RPI, assuming we maintain that).

North Dakota (W)
RPI gap: +.0095
TUC: 14-15 vs. 15-13
Common: Air Force (both 1-0), Boston College (both 0-1), Minnesota (1-0-1 vs. 1-3), Michigan Tech (1-0 vs. 1-0-1). This is a lock for us.

Keep RPI and we've got it.

Merrimack (L)
RPI gap: +.0122
TUC: 14-15 vs. 11-10
Common: Boston College (0-1 vs. 0-2-1). Can't win this.

Thanks to the stupidity of weighting each category equally regardless of the magnitude of the difference (and counting common opponents even when the number of games is 1), we have to make up the gap in TUC.

Cornell (W)
RPI gap: +.0131
TUC: 14-15 vs. 6.5-5.5
Common: None.

With no common opponents or head-to-head, RPI is all that matters and we're up significantly.

Notre Dame (W)
RPI gap: +.0231
TUC: 13-14 vs. 10.5-14.5
Common: -0.325
Head-to-head: Even.

The RPI lead is practically insurmountable (if we manage to blow that, we won't be anywhere near the bubble anyway); just have to keep TUC or swing common (and we can even afford to lose both if we play and beat them at the Joe).

Lake Superior State (W)
RPI gap: +.0286
TUC: 12.5-12.5 vs. 9.5-12.5
Common: +1.3
Head-to-head: -1

Only still in play because of head-to-head; even so, we would have to lose the TUC lead and either lose a huge edge in common games or lose to Lake State to lose this comparison.

Colgate (W)
RPI gap: +.0393
TUC: 14-15 vs. 6-8 (just barely a lead for us)
Common: Robert Morris (2-0 vs. 1-0), Ferris State (both 0-2), Miami (0-2 vs. 1-1)

A sweep of Miami would clinch this by making common opponents no worse than a tie (as would a 2-1 series win and a win over Ferris at the Joe). TUC record could swing this the wrong way if we don't get either of those.

New Hampshire (W)
RPI gap: +.0440
TUC: 14-15 vs. 9-15
Common: Boston College (0-1 vs. 0-3)

Only a potential danger if UNH beats BC in a tournament game, and even then it would take a massive TUC swing.

Massachusetts (W)
RPI gap: +.0463
TUC: 14-15 vs. 9-14
Common: Boston College (0-1 vs. 2-1)

Similar to UNH, except they've already gotten the "beat BC" part taken care of. Have to keep TUC record.

Teams Near the Cut Line

Seeing UNH and UMass drop out would be nice just to avoid the distant possibility of a TUC-record swing and single-game common opponent comparison stealing those points from us. Michigan Tech (.0038 below the cut line) and Alaska (.0069) moving up would boost our TUC record. (Robert Morris making the jump would also be nice, but they're too far back to realistically believe that will happen.) We would also like to keep Air Force in.

Rooting Interests for the Week

St. Cloud State over Minnesota-Duluth: SCSU is too far back to catch us in RPI, but we could catch UMD and steal the point that way.

Northeastern over Boston University: The small risk of NE swinging TUC (which requires us to lose to Miami anyway) is nowhere near enough to offset potentially catching BU in RPI.

Wisconsin over Minnesota: Won't help them enough to overtake us in common opponents, but it will let us pass Minnesota in RPI.

New Hampshire over Maine: We can safely lose TUC to UNH if they don't face BC again (and it would take more than just this anyway); this would give us TUC over Maine.

Bowling Green over Northern: It won't hurt Northern's TUC record, but it would kill their RPI and common opponents. This is probably less important than the others listed so far.

Nebraska-Omaha over Denver: Extra RPI gap and closing the gap in TUC would be helpful. The RPI boost for UNO is harmless.

Minnesota State (Mankato) over North Dakota: Expand the RPI gap.

UMass over Merrimack: This would give us TUC on Merrimack, although it would bring UMass closer to taking TUC on us. The other way around is a lower-variance option (more likely to lose to Merrimack and beat UMass).

Ohio State over Notre Dame: ND's a serious threat in TUC; Ohio State is a mild threat in RPI and TUC but our head-to-head edge should keep us safe as long as we advance to the Joe.

Alaska over Lake State: Really want to get Alaska under consideration and, if possible, knock Lake State out.

Michigan Tech over Colorado College: Getting Tech under consideration would help our TUC record.

Another update will be posted next week.