#6 UND has moved much closer to clinching a tournament berth. Winning at least 2 of the remaining 6 would have UND more likely than not in the top 13 in PWR going into the conference tournaments. Even if UND won only 0 or 1 of its remaining games, an at large berth would still be within reach with a good showing in the conference tournament.
Given the above, UND faces a somewhat predictable situation this week of being able to make a small gain with a sweep over Denver, stay pretty much the same with a split, or fall a bit more if swept.
The final weeks of the regular season
Only three teams can’t fall from being a TUC by the end of the regular season: #1 Quinnipiac, #2 Minnesota, and #3 Miami.
Quinnipiac (PWR Details) is particularly safe, even a catastrophe seems to only drop them 3 spots. Winning only 2 of their remaining 4 would leave them with only about a 10% chance of falling from the #1 spot. Their .7778 vs. TUCs is unassailable, and .5828 RPI is miles ahead of #2 Minnesota .5658. To put that RPI into perspective, if Minnesota swept its final 6 games their RPI would only rise to about .5738 (Minnesota RPI details).
#2 Minnesota and #3 Miami each also have good RPI leads over the teams chasing them, coming in at .5658 and .5529 respectively, with #4 New Hampshire at .5477.
However, #2 Minnesota and #3 Miami aren’t alone in vying for the #2 PWR ranking at the end of the regular season. A staggering 9 teams could claim that position: #2 Minnesota, #3 Miami, #4 New Hampshire, #5 Boston College, #6 North Dakota, #7 MSU-Mankato, #9 St. Cloud St, #11 Denver, and #14 Mass.-Lowell.
Interestingly, the list of teams that can finish top 4 isn’t much longer, add only #8 Western Michigan and #10 Niagara to the list.
The team with the most upside potential for the remaining regular season is #29 Providence, which can climb to #8. That’s mostly just because big upward moves are possible from that low a rank (#28 Colgate and #30 Colorado College could each rise to #14).
The team with the most downside potential for the remaining regular season is #13 Boston University (BU PWR details), which could fall to not being a TUC. #13 BU is only #20 in RPI, and is already losing most of its common opponents comparisons.
Methodology
Each forecast is based on at least one million monte carlo simulations of the games in the described period. For each simulation, the PairWise Ranking (PWR) is calculated and the results tallied. The probabilities presented in the forecasts are the share of simulations in which a particular outcome occurred.
The outcome of each game in each simulation is determined by random draw, with the probability of victory for each team set by their relative KRACH ratings. So, if the simulation set included a contest between team A with KRACH 300 and team B with KRACH 100, team A will win the game in very close to 75% of the simulations. I don’t simulate ties or home ice advantage.
Resources
- Current PWR Rankings (SiouxSports.com)
- Current RPI Rankings (SiouxSports.com)
- CHN PWR Rankings (CollegeHockeyNews.com)
- TBRW build your own rankings (slack.net)
- Explanation of how PWR mimics NCAA tournament selection (CollegeHockeyNews.com)
A question came in via email “why does RPI seem more affected by losses than wins”. Here was my response:
If you’re looking at a high RPI team, like North Dakota, it’s behaving a lot like averages.
Say you’re keeping a win/loss record instead (RPI is just a fancier
version that includes opponents win/loss, but it will behave the same
mathematically).
If after 10 games you’re 7-3, your win% is .700.
Win the next one, you’re 8-3, and your win% is .720.
Lose the next one instead, you’re 7-4, and your win% .636.
Because our starting point was closer to 1 than 0, averaging in another 1 moves our average less than averaging in another 0.
Similarly, you can think of each game as having an RPI value for winning
and an RPI value for losing (and in fact, each does, and you can view them on a team’s RPI details page on this site). Whichever value your current RPI is closer to
will make your RPI move less than the one you’re further away from.
Let me go a step further than I did in that email reply and give an actual example:
UND RPI details
UND’s RPI is .5457
The games against Bemidji St next weekend are a good example. The game RPI value of a win is .5807, while the game RPI value of a loss .3307.
Even if UND sweeps, averaging in two .5807’s with UND’s other 32 games won’t move UND’s RPI much at all. Averaging in two .3307’s could actually pull it down a noticeable bit.