Michigan Hockey now # 1 in Pairwise
January 22nd, 2022 at 11:03 PM ^
The narrowest of margins: We are 5 ten-thousanths of a point ahead of Minnesota State in RPI, after Minnesota State beat St. Thomas, which is a much worse team than Minnesota. That appears to be the difference.
Our common opponent winning percentage is identical, and we have the head-to-head win, but when Minnesota State leads in RPI (as they have for much of the season) RPI is the tie breaker.
January 23rd, 2022 at 10:10 AM ^
I had season tix to RPI hockey back in the late 80’s/early 90’s. Pretty solid squads back then.
Oh, not that RPI
January 23rd, 2022 at 12:21 PM ^
I also had season tickets for RPI, 91 through 96. Did the "hockey line" too. They had some great teams, and in 95, we went up to Lake Placid for the ECAC Championships. That was probably my best college hockey experience.
I went to games and Freakouts up until like 2012, but they were pretty consistently awful after their 80s/90s best years.
January 23rd, 2022 at 1:13 PM ^
season tix from 82'- about 92.
Oates, Dark, Puppa, Carter, thru Juneau and Coles. Try to hit a game or two every year but obv. they have closed for the last 2 seasons.
January 23rd, 2022 at 5:02 PM ^
You had the best era for sure. I loved '95, but '85 would have been epic. In the eighties, the only thing I knew about RPI was they had something called a "Fieldhouse" where the WWF would go sometimes. Then they threw a bunch of money at me in 1990 and that was that. I enjoyed much of my time there, and hockey games were some of the most fun experiences.
January 23rd, 2022 at 6:23 PM ^
Yeah, Mike Tyson fought there once or twice too. Before he won the belts
January 23rd, 2022 at 6:23 PM ^
January 22nd, 2022 at 11:12 PM ^
I admit the following is a duplicate from another thread (with RPI values updated and cursing deleted). If my copy/paste job is a miscalculation and no one cares or agrees about the opacity, please neg my self-absorbed self to Bolivia.
I saw that, too. But it's by the skin of our teeth (our RPI is now #1 at 0.6095 to Mankato's #2 at 0.6090). And RPI is the tie breaker.
Having said that, I still need an explanation on why we need the tiebreaker. We have the H2H (= 1 pairwise point) ... and by my accounting, we should also have the "record vs common opponents" advantage (*should* = 1 pairwise point). But USCHO and CHN deny this.
Maybe that's reality but if so it's super opaque. WHY ON EARTH is Mankato's 7-1-0 considered numerically equal to our 5-0-1?!? What's the arithmetic there? I can see 7-1-0 easily equaling 0.875. But what is so incorrect about this mathematical operation for Michigan: (5 + 0.5)/6 = 0.9167 ??????
The only way our common opponent record (5-0-1) equals 0.875 is if our tie is only worth 0.25. But why the heck would that be the case? Apparently, this is my new hobby horse. I find the lack of clarity around this so weird.
January 22nd, 2022 at 11:33 PM ^
That's a good question. I'm going to do my best Alton impression (he is the best there is at this kind of thing) and go to the USCHO explanation page where we find the answer:
Starting in the 2011-12 season, the common opponent calculation compares the sum of the winning percentages against each opponent
So it's not actually the winning percentage of all the games against common opponents put together. It's a question of how you do against each of the opponents you play in common. So while Michigan has played two fewer games against common opponents than Mankato (one game each against Michigan Tech and UMD, instead of Mankato's 2 each) neither Michigan nor Minnesota State are either punished for or benefit from the differential in games. Our .500 record against Tech for the single tie counts the same as Mankato's 1-1 record against LSSU.
So we get 1.000 for our 2-0 record against UMass, .500 for the Tech tie, 1.000 for the single win against UMD, and 1.000 for the 2-0 sweep against LSSU. Minnesota State gets 1.000 for its 2-0 record against UMass, 1.000 for it's 2-0 record against Tech, 1.000 for its 2-0 record against UMD (no extra benefit for the extra game in this case), and .500 for it's 1-1 split with LSSU.
Both get a common opponent total of 3.500. So it's even.
Minnesota State has two more games against Michigan Tech, and a non-perfect record would knock down their common opponent record. Their games against BG do not count, since our game against them was an exhibition. And there are still playoff games to play; none of our common opponents are in the B1G, but we've played two of their potential playoff opponents. Minnesota State could improve their percentage against LSSU if they played them, but they can only mar their record against Tech.
January 22nd, 2022 at 11:51 PM ^
That's illuminating. Thank you!
Now, do you find it strange that CHN reports 0.875 = 0.875 for U-M's and MnSU's COp comparison? Your description sounds correct (unfortunately), but it also sounds incompatible with CHN's arithmetic.
January 22nd, 2022 at 11:58 PM ^
I just figured that out: it's a single-number calculation of the accumulated winning percentages divided by the possible total winning percentage. 3.500 divided by a possible total of 4.000. Why that's what they show is unknown to me, but that explains the number.
January 23rd, 2022 at 12:11 AM ^
Daaaaamn. That's actually it. I wasn't nearly open-minded as I believed about this and missed it, getting too distracted by the 7-1-0 = 0.875 coincidence. You're freaking good. And the COp calculation is still wickedly unintuitive.
For the next topic, your previous point about MnSU having future dates with Michigan Tech is key. Given our SOS comparison for the rest of the season, I very much doubt we'll finish with a higher RPI than Mankato. Unless we fall apart, it seems realistic to predict their end-of-February match-up with Tech is going to be decisive for the #1-overall seed.
A split MnSU/Tech split would win us back COp and thus the comparison, right?
January 23rd, 2022 at 12:31 AM ^
I think so, assuming they do not beat LSSU in the playoffs; haven't calculated that far and I'm not going to bother. If they drop the common opponent comparison, they cannot beat us even with superior RPI.
I couldn't figure out the CoP issue at first, but I assumed that there was probably a good reason for it to look like that, and did the math and there it was.
I guess I sort of understand why the committee uses the cumulative win percentages. They're comparing teams based on how they fare against common opponents, but what they're actually looking for is, how many common opponents are they better than, and by how much? If, say, you have three common opponents, and you are better than your counterpart team against two of them, and the third is the best in the country and you're both winless against them, is it fair the other team plays them once and you are in a conference with them and play them four times?
Or does it make sense to just say you're both winless?
It's complicated, but I at least get why they do it that way.
January 23rd, 2022 at 8:16 AM ^
Whenever I read detailed explanations of how college hockey determines team rankings I’m as lost as I was in my lone semester of calculus at U-M.
January 23rd, 2022 at 9:39 AM ^
Came for a hockey discussion, found the Michigan Math Nerds Society monthly statistics gala...
January 22nd, 2022 at 11:29 PM ^
So this means home ice during the playoffs, right?
/s
January 23rd, 2022 at 12:06 AM ^
Don't. Get. Me. Started.
January 23rd, 2022 at 9:41 AM ^
Start!! C'MON!!!!!
::sprays starter fluid into carburetor::
START, DAMN YOU!!!!!
January 22nd, 2022 at 11:39 PM ^
But are we going to have a good finish with 4 top guys in the Olympics?
January 22nd, 2022 at 11:50 PM ^
TBD. Still good players left on the roster, including our goalie. And our RPI is pretty durable; as it is, we have more losses than a lot of teams below us, but our strength of schedule is excellent thanks to non-conference games against Western, UMass, UMD, and Minnesota State among others. Michigan has played 8 of the top 12 teams and is itself the 9th such team. So unless there's a real collapse with those players gone (and I won't rule out some bad games) the team should be pretty secure in tournament position. Losing the top seed seems... likely to me. Losing a #1 seed isn't impossible, but that will require either Quinnipiac or Saint Cloud to jump Michigan in RPI, in addition to the other top four teams; Michigan's position against UMass and UMD is basically unassailable thanks to H2H and common opponent records.
January 23rd, 2022 at 8:13 AM ^
January 23rd, 2022 at 11:38 AM ^
Luckily, we only have two series while the players are gone (assuming nobody gets CoVid).
We have Wisconsin next week and as of now, we expect to have everyone (Canada MAY have a training camp before announcing final roster). First week of February we have an exhibition against the USNTDP Program which won’t count. The four are expected to miss the MSU and OSU series. We should have enough talent on hand to be fine in the MSU series. We should still be the better team there, but nothing is guaranteed in hockey. But getting at least a split against OSU is huge both for Pairwise and for B1G standings.
Pairwise is hard to predict because it is based so much on what other teams are doing in addition to your own team. But as long as we don’t go 0-4 while the four are gone, I wouldn’t be overly concerned about not getting a 1 seed. There’s still the Wisconsin series, the ND series, and the B1G Tournament to make up some ground with 2-3 losses.
Nightmare scenario would include a split (or worse) with Wisconsin (which would be a real drag on our PWR with Wisconsin at #37) and going 0-4 with the four players gone. Losing 2 to MSU (#31) would be a bummer and also be a significant drag.
Following the Olympics, we only have the ND series and the B1G tournament to make up any lost ground. So sweeping Wisconsin and avoiding going 0-4 while the players are overseas would be huge and make us virtual locks for a 1 seed.
January 23rd, 2022 at 4:25 AM ^
Michigan is becoming increasingly safe for a 1-seed.
Prediction: if the tournament committee can put WMU in our regional as a 2-seed, they will. They’re a tough match-up for anyone, period, maybe a particularly tough one for us on the basis of age/style/physicality. And the intangibles of the over-dramaticized GLI cancellation could only add pep to their step.
January 23rd, 2022 at 9:45 AM ^
Hockey committee doesn't really do that. It does do things based on geography and to avoid conference matchups and for that reason Western might be with us if they drop a little. Because right now Western is tied with Denver for the fourth #1.
Today's bracket would probably be:
#1 overall Michigan going to Allentown, Pennsylvania
#2 Minnesota State to Loveland, Colorado
#3 Denver to Albany
#4 Western to Worcester
The two seeds are UMass, St. Cloud, Quinnipiac, and Duluth (in that order). Denver, Western, St. Cloud, and Duluth are all in the same conferences so they don't get the same regionals. So that means that St. Cloud and Duluth have to go to Michigan and Minnesota State's regionals. UMass goes to Worcester, Quinnipiac to Albany, Duluth to Allentown (joy), and St. Cloud to Loveland.
Three seeds are Tech, OSU, Minnesota and Notre Dame. So we obviously get Tech to avoid conference games. Minnesota goes to Colorado for geography reasons. Notre Dame to Massachusetts and Ohio State to Albany.
Last teams in are North Dakota, and then Merrimack and UMass Lowell. But Providence is tied with them in Pairwise rankings (lose RIP tiebreaker). And finally you get the crappy conference champion, American International. We get them as the #1 overall seed. Then North Dakota has to go to Loveland to avoid a conference game in round one. Merrimack and Lowell are whatever, but we sent UMass to Worcester, so we'll send Lowell to Albany.
So:
Allentown, PA
Michigan vs. AIC
Minnesota Duluth vs. Michigan Tech
Loveland, CO
Minnesota State vs. North Dakota
St. Cloud vs. Minnesota
Albany, NY
Denver vs. UMass-Lowell
Quinnipiac vs. Ohio State
Worcester, MA
Western Michigan vs. Merrimack
UMass vs. Notre Dame
Which seems fine, but that Loveland regional is fucking loaded and would be super fun.
January 23rd, 2022 at 10:35 AM ^
[This is great, thank you. After all that work, you might want to edit the bracket to include ND (as I know you intended), rather than listing OSU twice. (Hope it’s clear I’m not criticizing.)]
This bracket also underscores the value of the overall #1 as well as anything could. For this set-up, Allentown isn’t simply the least gladiatorial (fewest concentration of heavyweights); it’s by far least in that way.
Getting ahead of ourselves for fun: which 2-seed would be a best match-up for U-M?
Of the Western, UMd, SCSU, Quinnipiac, UMass multifecta, I want Michigan to play Western and SCSU least. A step down from those, I’d rather avoid modern UMd teams in every tournament for awhile, if given the choice. So I guess UMass seems the least threatening. But I’m also not intimidated by Quinnipiac. I think that (Quin’ac) could make for an interesting point of disagreement here, given their win%. I’m unaware of their style (prob an older team, though, right?) and likely biased against their conference and SOS, but whenever they get #1 votes, my gut feel is “holy paper tiger.”
January 23rd, 2022 at 4:29 AM ^
The best way to avoid a rough regional (any match-up) is to take the overall #1, but that will be a heavy lift. Maintaining our RPI will be a challenge without sweeps against the Big Ten dregs (MSU/UW) and a split vs OSU during the Olympic month (down Beniers, Power, Brisson, KJ).
Thankfully, we still have serious talent staying home in Bordeleau, Hughes, Samo, and Portillo. Clearly, the next month would be an ideal time for Beecher to solve and exit his slump.
January 23rd, 2022 at 7:58 AM ^
Let's just hope they become inconsistently much better going forward to where the team can win even when they play inconsistently.
I would love to see this team and Mel get a Hockey NC.
How sweet would that be? Pretty pretty sweet.
January 23rd, 2022 at 9:42 AM ^
Few things in life effectively blend beauty and utter stupidity as well as college hockey.
January 23rd, 2022 at 11:16 AM ^
Comment of the year so far.
January 23rd, 2022 at 10:25 AM ^
They are playing much more consistently now. Will be a team to be reckoned with in tourney.