[Bumped. Fantastic stuff. Beat Michigan. -KJ]
Michigan State vs. Michigan
Friday, Saturday, and Sunday (if necessary) at Munn Ice Arena
No upsets in the CCHA first round (in fact, the home team swept all four series), so Michigan comes to town. This preview is going to be a bit light on personal observations; I've seen very few games this year (one of the hazards of living over 1000 miles away) and thus have to rely on the boxscores and team stats. Pairwise update will follow the series preview.
Tale of the Tape
Season series: MSU won 3-1 (3-2 at Yost, 2-0 at Munn, 3-2 at Munn, 4-5 at Joe Louis)
Despite a better record, we're both scoring fewer goals and allowing more overall. In conference play, our goals against was slightly better but net goal difference was still worse. If we define difference between the record you would expect based on goal difference and the actual record as "luck", then Michigan has been spectacularly unlucky this year: one game over .500 in conference despite a +14 GD in 28 games is pretty hard to do. My minimal observations, combined with reading some previews and recaps from Brian and looking at the boxscores, have led me to a possible alternate explanation for this phenomenon. (Be sure to check there, as his preview will almost assuredly be more thorough than mine. That's the advantage of actually being able to see games.)
One observation Brian had early in the season is that Michigan has had a tendency to put up insane shot totals but lose as most of the shots are stopped and the few shots the opponents get trickle through. In losses this season, they have outshot opponents by 10, 13, 15, 12 (the 2-0 loss at Munn), 10, 25, 18, 33 (a 4-3 GLI semifinal loss), 15 (their only tie of the season), 9, and 18. Their 33 shots on goal per game is probably near the top nationally (I can't find a list of teams ranked by shots per game), but they rank just 20th nationally in goals. This leads me to suspect that Michigan tends to just throw the puck at the net at every opportunity and hope something good happens - not a bad strategy in general. Against bad teams, these chances typically result in a lot of goals as they get many good scoring opportunities or pick apart a weak goalie. But against stronger opposition, this often leads to frustrating games where you're outshooting the opponents by a wide margin but they're all poor-quality shots and easy saves. (Red Wings fans will probably recognize that description all too well.)
Game-by-game goal totals seem to bear this theory out a bit:
(We have one game scoring more than 6 - a 10-1 win over Michigan Tech. Rather than stretching the axis, I just lumped that into the "6+" category.)
Michigan's per-game totals against all opposition have a surprisingly flat distribution. Break it out into games against the RPI top 25 and the bottom 31, and there's a clear pattern: they scored 3 or more goals 80% of the time against the lower tier, but just 37% against the top. We, on the other hand, scored 3+ 67% against the lower tier and 56% against the upper. Michigan does have four 5+-goal games against teams under consideration (Alaska, Minnesota, Northern, and us) to our one, but they've been shut out twice by TUCs and held to one goal four more times. We've been held to one goal or none six times against TUCs as well, but we put up 3+ ten times to Michigan's seven.
What does it all mean? Our defensemen and goalie Drew Palmisano are going to have their work cut out for them. An off night in our own zone is probably going to be an ugly loss. If we can limit Michigan to taking mostly low-quality shots, we may well get outshot but we'll likely come out ahead in the only category that matters. Palmisano is certainly capable of stealing a game or two if we get outplayed; he ranks ninth nationally with a .922 save percentage. Michigan's goalie situation is not so bright: Bryan Hogan ranks 55th at .901. (Hogan has been out for the last few games due to injury; backup Shawn Hunwick is .915 in five appearances, which would be good for 18th if he had the minutes to qualify. He has two shutouts - one of which was the game in which Hogan was injured midway through the first - but gave up five to Notre Dame on just 25 shots in another game. I haven't been able to find any confirmation on Hogan's status for this weekend.)
A fair amount of movement around us over the weekend. New Hampshire's 0-1-1 weekend against Boston College has us stealing the point thanks to a tiny TUC lead (9-9 to 11-12). Unfortunately, UMass sweeping Maine turned out to be far more significant than I expected: it gave them an insurmountable common opponents lead and a lead in TUC (11-9 to 9-9). End result: Still 12th, still with 12 out of a possible 24 points. Assuming Bemidji State wins the CHA tournament and the four major conferences (WCHA, CCHA, ECAC, Hockey East) do not have a surprise auto-bid winner, we need to remain at least 15th to get in.
Doesn't look likely that there will be much movement around the TUC cliff. There's a .006 gap (which works out to about the equivalent of a win for the trailing team and a loss for the leading team, give or take) between RPI #25 BU and #26 Northeastern, and Northeastern's season is done. Ohio State is another .0027 back and about to be disemboweled by Miami. Short of a Merrimack sweep of BU or a major upset or two involving teams near the cut line, the top 25 looks unlikely to change.
Comparisons in play:
- Alaska: Their RPI lead has grown to .0066. We lead TUC, 8.5-7.5 to 7.5-7.5 (head-to-head does not count in TUC) and common opponents 17.5-9.5 to 17-13. 0-1-1 head-to-head. If we don't face them at the Joe, we have to flip RPI. If we do, we have to beat them, as there's no way we can overcome the RPI deficit otherwise; playing them even is probably enough the rest of the way, as we'll only need to either keep TUC or flip RPI.
Last week: 2-0 vs. Western Michigan 4-0, 4-1
This week: CCHA quarterfinals @ Northern Michigan, best-of-3
- New Hampshire: Their RPI margin has dipped to .0057, and we've overtaken them in TUC, 9-9 to 11-12. Common opponents is a lock for us. Keep TUC and this is ours. I don't see any route for us to flip RPI while losing TUC (unless Michigan drops out of the top 25, and that appears unlikely, though not impossible, even if we sweep them), but if we do that we win the comparison as well.
Last week: 0-1-1 @ Boston College 3-3, 2-3
This week: Hockey East quarterfinals vs. Vermont, best-of-3
- Cornell: Our RPI deficit has grown slightly to .0031. Cornell is still short of the 10 TUC games needed for that criterion to count, at 4-5 to our 9-9. Common opponents is a lock for us. Win our series with Michigan, and assuming Michigan stays in the top 25 we guarantee the comparison if Cornell faces Union and/or Yale in the final four and loses or splits (even if we lose both at the Joe, we'll be 11-12 at worst and they'll be 4-6 or 5-6). If Cornell beats both to win the ECAC tournament, we can't take the comparison without winning the CCHA tournament (we won't be able to catch them in RPI, and the best we can do in TUC would be a tie at 12-10 to 6-5). If Cornell beats one and does not face the other, we just need to beat Michigan and avoid losing both games at the Joe (even a tie in the third-place game, assuming that it would be decided via shootout to avoid delaying the championship game, should be sufficient if we sweep Michigan; if we win in three we would have to win one of the two games at the Joe). And if they don't make it to the final four, a series win over Michigan probably flips RPI (though we might need to avoid losing twice at the Joe).
Last week: Idle
This week: ECAC quarterfinals vs. Harvard, best-of-3
- Ferris: Same situation as last week. If we meet at the Joe, winner probably takes the comparison - although if the loser wins their series in two and wins the third-place game while the other wins their series in three and loses the final the head-to-head point could be overcome (we might flip RPI in that scenario, and they would flip common games while keeping RPI). If we don't meet there (which requires that one of Ferris, Miami, or us gets upset in the quarters), matching Ferris win for win keeps TUC and common opponents and gives us the point.
Last week: Idle
This week: CCHA quarterfinals vs. Nebraska-Omaha, best-of-3
- Northern Michigan: Same situation as last week. If we meet at the Joe, winner probably takes the comparison, although scenarios like the ones for Ferris could overcome that. If we don't meet there, we need to keep RPI and common opponents, which means matching them win for win. An extra win might be necessary, even, because Northern is so close in RPI and will face teams with higher RPIs than we will in the quarterfinals (and, should we both make it there, the semifinals).
Last week: Idle
This week: CCHA quarterfinals vs. Alaska, best-of-3
- Minnesota-Duluth: We've gained a little ground in RPI (the gap is down to .0012). We lead TUC, 9-9 to 9.5-12.5. They lead common opponents 9-5 to 5.5-3.5 (possible relevant opponents for us: Northern; possible relevant opponents for them: Minnesota, Wisconsin, Michigan Tech). If we keep the TUC lead and defeat Northern at Joe Louis (only way we can meet in the semis is if Ferris is upset by UNO and the rest of the chalk prevails), UMD would have to go 1-0 or 2-0 (1-0-1 is sufficient if they win against Tech or Minnesota in the opener of the Final Five, lose to someone else in the semis, and tie the third-place game assuming that's allowed) against relevant opponents to keep common opponents. If we do not face Northern, we take common opponents only if UMD goes 0-1 or 0-2 (or 0-1-1) against relevant opponents. If we lose to Northern, we cannot take common opponents and would need RPI to win the point.
Last week: 1-1 @ Alaska-Anchorage 2-3, 6-3
This week: WCHA first round vs. Colorado College, best-of-3
- Nebraska-Omaha: RPI lead is down to .0045. Down in TUC (7.5-8.5 to 9-9), up in common opponents (15.5-10.5 to 16-12), 1-0-1 head-to-head. Keep the RPI lead and we just have to keep common opponents (very likely if we keep the RPI lead). Lose that and we need to make TUC even at worst and keep common opponents (unlikely if we lose the RPI lead). Should we meet at the Joe, it will be a similar scenario to those for Ferris and Northern - winner probably gets the point unless everything else goes as far the other way as possible.
Last week: 2-0 vs. Bowling Green 6-1, 6-1
This week: CCHA quarterfinals @ Ferris, best-of-3
- Vermont: We lead RPI by .0060. Common opponents is now guaranteed. We have a slim lead in TUC, 9-9 to 10.5-11.5. Keep either RPI or TUC and we win the point.
Last week: 0-1-1 at Mass-Lowell 2-5, 1-1
This week: Hockey East quarterfinals @ UNH, best-of-3
- Mass-Lowell: Our RPI lead is still a solid .0105. Common opponents is a tie at worst (that only if Mass-Lowell sweeps Maine and we lose to UNO at the Joe). TUC record is ours as well for the moment: 9-9 to 8.5-10.5. Pretty hard to lose this if we win the series with Michigan, although it might be possible if we lose twice at the Joe and Lowell wins the Hockey East tourney.
Last week: 1-0-1 vs. Vermont 5-2, 1-1
This week: Hockey East quarterfinals @ Maine, best-of-3
- UMass: The poster child for "Why The Pairwise Is Weird". We lead RPI by a whopping .0141, but we lose the comparison because they're 4-0 vs. common opponents (3-0 against Maine, a win against Rensselaer) and lead TUC 11-9 to 9-9. We have to flip TUC. Hopefully Boston College will do us a favor there - a sweep means that as long as Michigan stays in the top 25, getting to Joe Louis will do it for us if we don't beat Michigan in three and then lose both at the Joe.
Last week: 2-0 at Maine 5-2, 4-3
This week: Hockey East quarterfinals @ Boston College, best-of-3
- Maine: RPI lead of .0140 is practically insurmountable (barring a collapse which would have a lot of other teams passing us anyway). With no common opponents, RPI is enough to keep the point.
Last week: 0-2 vs. UMass 2-5, 3-4
This week: Hockey East quarterfinals vs. Mass-Lowell, best-of-3
- Colorado College: Our RPI lead is just .0010. TUC record looks like a safe win. We only flip common opponents if CC loses to two of Minnesota, Wisconsin, and Michigan Tech in the WCHA Final Five. Best bet is to keep the RPI lead.
Last week: 1-1 in home-and-home vs. Denver: 2-1 away, 3-7 home
This week: WCHA first round @ Minnesota-Duluth, best-of-3
- Union: RPI lead is .0110. Thanks to UMass replacing Northeastern in the top 25, Union has nine TUC games instead of eight now. We lose common games, but it will be hard to lose TUC now even if Union beats Cornell and Yale to get to 5.5-5.5. This one's safer than it was last week simply because of UMass making the top 25 (otherwise Union would be 5.5-4.5 if they win the ECAC, which would beat us unless we win the CCHA tourney).
Last week: Idle.
This week: ECAC quarterfinals vs. Quinnipiac, best-of-3
- Boston University: RPI lead is .0142. Dead even in TUC (9-9 to 10.5-10.5), slight lead in common opponents (5.5-4.5 to 3-3; their only potential relevant opponent is Maine, ours is Michigan). Win the series against Michigan and we lock up common opponents, which with our RPI lead makes this a virtual lock. Get swept, and we need BU to lose to Maine in the Hockey East final four. Lose 2-1, and we need them not to beat Maine (they can lose, tie, or not play).
Last week: 2-0 in home-and-home vs. Northeastern: 5-4 home, 4-3 away
This week: Hockey East quarterfinals vs. Merrimack, best-of-3
We seem to be in reasonable shape as teams behind us start having to play each other. That limits the amount of damage that can be done. With Vermont and UNH facing each other, and Ferris and UNO, it's unlikely that any more than two of those teams can take the point if we lose 2-1.
All predictions here should be taken with a grain of salt, as there are so many games in play and it's hard to say exactly how much the RPI will swing one way or the other on a win or loss, but the bottom line: Win the series and it will be hard for us to be left out (though it might still be possible if we lose twice at the Joe). Lose 2-1 and we're likely to be sitting right around the cut line. Get swept and all bets are off.
Conference Tournament Rooting Interests
All with an eye toward guaranteeing qualifying. Seeding really doesn't make that much difference except on the 1 line, where you are potentially playing for a home site and/or a weak auto-bid for your first opponent. Otherwise, bracketing follows the S-curve perfectly as long as it doesn't result in a conference matchup in the first round.
- #7 Michigan at #2 MSU: Duh.
- #5 Alaska at #4 Northern: Alaska winning this is the low-variance option: Northern almost assuredly won't catch us (unless they lost 2-1 and we get swept) - the difference between playing Alaska and playing Michigan shouldn't be enough of an SOS boost for them to flip it - and we'll be hard-pressed to catch Alaska even if we win. If we beat Michigan, a Northern series win might let us catch Alaska but it keeps Northern in play as well. A bad showing at the Joe might leave us losing both, a split there with Northern not winning it all and not beating us may let us win both. If we lose to Michigan, a Northern series win all but guarantees we lose both. Importance: Critical if we lose, significant if we win
- #8 Ohio State at #1 Miami: An upset here means we would play OSU first and be a (slight) favorite to win the auto-bid, but this really doesn't matter much. OSU is the only non-TUC still alive in the CCHA tourney (assuming Michigan stays in the top 25). The only complicating issue is the potential for bid-thievery. Better to have Miami (who can also help knock off the other CCHA bubblers - we won't face them until the final or third-place game) and not risk a stolen bid. Importance: Minimal
- #6 Nebraska-Omaha at #3 Ferris State: Neither team is far from taking the point on us. Root for a 3-game series. If we win our series, the loser here is not catching us. If we lose, we're losing the point to the series winner. If Ferris wins and we win, we play them (assuming Miami does not lose to OSU). If UNO does, we cannot face them in the semis but could in the final or third-place game. For matchup purposes, Ferris might be a better opponent to face than UNO should it come to that - we're 1-1 against Ferris, 0-1-1 against UNO. I don't know that we care too much who wins, as long as it goes three games. I lean slightly toward Ferris, but even bid-thievery isn't likely to matter here as either of these teams winning it all would push them into the top 15. Importance: Minor
Totally irrelevant, as they have no one in at-large contention.
Root for Bemidji State to win it all, as they're the only at-large contender in the conference. They face Niagara in one semifinal; Robert Morris vs. Alabama-Huntsville is the other. Importance: Critical if we lose, minor if we win
- #10 Michigan Tech at #1 Denver: Immaterial by itself, except in the unlikely case that MTU makes a bid-thief run. Best for Denver to win this and keep stomping everyone in sight, limiting any gains on us in TUC record by fellow bubblers from the WCHA. Importance: Minimal
- #9 Alaska-Anchorage at #2 Wisconsin: Ditto.
- #8 Minnesota State (Mankato) at #3 St. Cloud State: Ditto.
- #7 Minnesota at #4 North Dakota: Minnesota can't realistically catch us unless we fall apart (we have a solid RPI lead and an insurmountable TUC lead), but we're better off knocking them down to 1) prevent possible bid-thievery and 2) make it tougher for whoever emerges from the 5-6 series. Importance: Minor
- #6 Colorado College at #5 Minnesota-Duluth: UMD in 3 would be ideal. It keeps CC off our back (they're close in RPI and we would need a miracle to flip common games) while giving us a solid shot at taking RPI from UMD. Importance: Significant
- #5 St. Lawrence at #4 Colgate: Both would be bid-thieves if they win the whole thing. Importance: Irrelevant
- #11 Brown at #1 Yale: Prevent bid-thievery. Yale and Cornell are the only sure bets for the tournament from the ECAC at this point (and even Cornell could fall out if they get swept and certain other series go the right way). Importance: Significant
- #7 Quinnipiac at #3 Union: Union could play their way in (and has an outside chance of stealing the point, especially if UMass should fall out of the top 25). Quinnipiac is auto-bid or bust. Easier road for the teams that are definitely in makes bid-thievery less likely. Importance: Minor to significant
- #9 Harvard at #2 Cornell: Harvard winning this series almost certainly gives us the point on Cornell (if we win the Michigan series), but it also increases the risk of bid-thievery (Yale would be the only team left assured of a spot already). Given that we would pass Cornell in the pecking order, that's a trade I'm more than willing to make. Importance: Significant
- #8 Vermont at #1 New Hampshire: If we win, we want this series to go 3. If we lose, we want it to be a sweep. A UNH win is probably better; it risks losing that point on TUC record but we'd be losing the point to someone who's already (relatively) safe rather than risking a loss of the Vermont point to a fellow bubbler. Importance: Minor to significant
- #5 Mass-Lowell at #4 Maine: Maine can't catch us barring an RPI disaster, Mass-Lowell is a little closer in RPI but would also need to flip TUC or tie common opponents (the latter only possible if they sweep this series). The key point is probably that Mass-Lowell is ahead in the Pairwise (winning eight comparisons to Maine's five). But I think we're okay with anything that could happen here as long as we win our series; if we lose, Maine is the safer point but not by much (and we may need them to deal with BU; see below). Importance: Significant if we lose, minor if we win
- #6 Merrimack at #3 Boston University: Win our series and this is moot. Get swept, and we lose the comparison to BU (assuming they don't drop out) unless they win the series and then lose to Maine. Lose 2-1, and we're mostly safe (though if BU plays Maine they must not win). Importance: Irrelevant if we win, minor if we lose 2-1, major if we lose 2-0
- #7 UMass at #2 Boston College: Eliminate a potential bid thief, and knock UMass's TUC record into range for us to prevent them from stealing the point. Importance: Significant to major
After this weekend's games, USCHO will put up the Pairwise Predictor, which lets you input hypothetical results for the final weekend and see the resulting Pairwise standings. So by Sunday night we should know exactly where we stand and what we need. Win the series and we'll probably be looking more at seeding and bracket possibilities than whether or not we're in. Lose and it will be a long wait until the tournaments resume next weekend and we learn our fate.