Monthly Archives: February 2016

This weekend’s big PWR movers

Leave a reply

Looking up!

As mentioned in yesterday’s PWR outlook article, all the teams from #20 St. Lawrence to #26 Clarkson have the potential to climb into the teens this weekend. But, #25 Minnesota-Duluth and #26 Clarkson, in particular, have the potential to make huge jumps.

The #25 Bulldogs could find themselves as high as #14, a jump of 11 spots, with a sweep of #2 St. Cloud St. Most likely with a sweep is a still unusual climb of 8-9 positions to #16-17.

duluth

#26 Clarkson could find itself as high as 16 with wins over Dartmouth and Harvard, though #19-21 are more likely. The likely outcomes for Clarkson with two wins cover a surprising #16-#23 span. Clarkson’s outcome with two wins would rest a lot on other games, with the Golden Knights cheering for sweeps from Alaska Anchorage over Minnesota St, St. Cloud St over Minnesota-Duluth, Maine over Northeastern, Alaska over Michigan Tech, and Canisius over Robert Morris.

clarkson

Going down?

Every team in the #14-22 range, except idle #15 Penn State, could fall more than five positions with a pair of losses. #14 Cornell, #17 Miami, #18 Michigan Tech, and #16 Minnesota are facing particularly steep falls.

Cornell could see itself fall as low as #24 with losses to Rensselaer and Union, though the spread of likely outcomes stretches a huge span of 17-24.

cornell

Miami faces a steep drop to as low as #29 if swept by Colorado College, with a still significant drop to 26-27 most likely.

miami

Michigan Tech could find itself as low as #28 if they drop a pair to Alaska, though #24-26 are more likely.

mtech

If Minnesota gets swept by Michigan, it will face a broad range of potential outcomes that stretches as low as #25. #19-23 are all reasonably likely in that scenario.

minnesota

Last full weekend of the regular season PWR outlook

Leave a reply

Half the conferences end regular season play this weekend, so I’ll mostly be taking a short-term focus for a couple weeks as we transition from the regular season to conference tournaments. I will, however, comment on the long-term implications when possible1.

Who’s in?

The top 5 are starting to look like locks for the NCAA tournament. It would take catastrophic results, really bad luck, and a lot of conference tournaments being won by low PWR teams for any of #1 Quinnipiac, #2 St Cloud St, #3 Boston College, #4 North Dakota, or #5 Providence to miss out on an at-large bid. #6 Michigan is just too close and has too many games remaining to make the call yet.

quendofseason

scsuendofseason

bcendofseason

undendofseason

providenceendofseason

michiganendofseason

A new #1 this weekend?

Despite winning only half its games in 2016, Quinnipiac has stayed atop the PairWise Rankings throughout. If they drop a game this weekend against Brown or Yale, St. Cloud St and North Dakota remain positioned to take over if they can secure their own wins. Despite sneaking ahead of North Dakota by the slimmest of RPI margins, Boston College is not a threat to take #1 this weekend.

quinnipiac

Who’s safe this weekend?

The top 12 are generally secure this weekend, with no one particularly likely to fall below #14, regardless of the weekend’s results. Those include:
#1 Quinnipiac
#2 St. Cloud St
#3 Boston College
#4 North Dakota
#5 Providence
#6 Michigan
#7 Yale
#8 Denver
#9 Notre Dame
#10 Boston University
#11 Nebraska-Omaha
#12 Harvard

harvard

Action on the bubble

#13 Mass.-Lowell through #19 Rensselaer, #21 Dartmouth, and #25 Minnesota-Duluth are all in play to come out of the weekend on either side of 14 (though climbing to 14 is mathematically possible for #20 St. Lawrence, it would take some luck). Those include:
#13 Mass.-Lowell
#14 Cornell
#15 Penn State
#16 Minnesota
#17 Miami
#18 Michigan Tech
#19 Rensselaer
#21 Dartmouth
#25 Minnesota-Duluth

masslowell

ren

dartmouth

duluth

The long shots

Teams between #20 St. Lawrence through #26 Clarkson (except those mentioned above) are fighting to break into the teens this weekend, positioning themselves for a climb to the bubble in coming weeks. Those include:
#20 St. Lawrence
#22 Northeastern
#23 Robert Morris
#24 Minnesota St
#26 Clarkson

stlawrence

clarkson

Who’s out?

It will be difficult for teams at #27 Bowling Green and below to move into position for an at-large bid.

bgsu

Methodology

Forecasts include the results of games played through Sunday of this week.

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

1Technical note—I can’t reliably use simulations to forecast across the line between regular season games, of which there are 87 remaining, and conference playoffs, which feature 31 best of 3 series with reseeding and 21 regular games. The total 1.5*10^51 scenarios is too many outcomes to exhaustively search, and sampling via monte carlos doesn’t work particularly well when the games’ participants aren’t known across the reseedings and elimination games.

This weekend’s big PWR movers

Leave a reply

As the weekend game results come in during the final few weeks of conference play, college hockey fans start watching the PairWise Rankings tables for big moves that might push teams into, or out of, position for an at-large bid. Here are the teams most likely to be cited on forums and in tweets this weekend for “surprisingly” large moves.

Don’t look down

#14 Penn State is unlikely to benefit much from a pair of wins over #40 Ohio State, but would likely plummet to about #20 (as low as #24 is realistically possible) if swept.

pennstate

Surprise! #15 St. Lawrence could climb just a little with a pair of wins over #10 Yale and #53 Brown, but is facing steep decline to the #22-24 range with a pair of losses (or a more modest decline with a single loss).

stlawrence

#17 Rensselaer won’t move much if they defeat #1 Quinnipiac and #55 Princeton, but face a steep decline of about 9 ranking positions if they drop both.

stlawrence

Going up?

#23 Minnesota Duluth, which I already noted is facing significant upside potential with a visit to #3 North Dakota, could surge to the #14-16 range if they hand the Fighting Hawks a second consecutive conference sweep. Even a single win would likely result in a noticeable climb.

duluth

#25 Robert Morris could excite fans with a likely move into the teens with a pair of wins over #30 Air Force. But remember, they’d need a repeat performance against Canisius next weekend to avoid giving that ground right back.

robertmorris

The mathematical long shots for at-large NCAA bids

Leave a reply

While #24 Minnesota St is pretty unlikely to climb to an at-large position by the end of the regular season, with a perfect run they could get close enough that an at-large bid wouldn’t be out of the question with some additional success in the conference tournament and a little luck.

The same is true, with decreasing likelihood with the lower ranks, down to about #29 Ferris St.

#24 Minnesota State
#25 Robert Morris
#26 Northeastern
#27 Bowling Green
#29 Ferris State

(Note that #28 Union getting into position for an at-large bid is particularly unlikely — that doesn’t mean mathematically impossible, it just didn’t come up enough in the simulations to be noticeable).

mankato

ferrisst

With two weeks until conference tournaments begin, teams on the bubble need to get hot

Leave a reply

Picking up where I left off yesterday, in A few top teams are starting to look like locks for the NCAA tournament, those teams currently ranked #10 and below in the PWR need to do some work to ensure being positioned for an NCAA tournament bid.

If #10 Yale and below lose more than they win, they risk going into conference tournaments in the teens and having to glue themselves to this blog. #11 Boston University and #12 Harvard have very similar outlooks to Yale, each with 4 scheduled regular season games remaining, and each wanting two wins to maintain their current position.

yale

#13 Mass.-Lowell can also maintain its current position, or slip slightly, with a pair of wins. But, that would leave them squarely on the bubble going into conference tournament play.

masslowell

Because the Big Ten regular season extends later into March, #14 Penn State has 6 games remaining. But, the Nittany Lions need 4 wins to make maintaining their #14 ranking the most likely outcome.

pennstate

#15 St. Lawrence has kept me in business this year, with a wild ride from #10 in my inaugural January article, down to as low as #27 at the end of January. I noted back in January that they would need to win about 9 of their last 10 to go into the conference tournament on the bubble. After a 5-0-1 run, the Saints are still positioned for a #13-14 ranking with no more losses, or a #15-16 ranking with one loss.

stlawrenceride

stlawrence

The #16 Minnesota Gophers have 6 games remaining, and most likely need 5 wins just to maintain that ranking going into the conference tournament.

That should be no surprise, as all the way back on January 6, I published a chart on Minnesota that suggested Minnesota needed to win about 12 of its remaining 16 to end up ranked #15-16. Including two additional tournament games, Minnesota has gone 8-4 over that period, leaving them little room for additional losses.

minnesota

#17 Rensselaer, #18 Cornell, #19 Michigan Tech, #20 Miami, #21 Dartmouth, and #22 Clarkson all need to win out to be most likely to end the regular season in the #13-14 range.

rpi

clarkson

#23 Minnesota-Duluth has a bit more wiggle room, with a finish in the #14-16 range likely with even just 4 wins in its remaining 6 games. The Bulldogs have the notable RPI improvement possibility, and challenge, of playing #2 St. Cloud and #3 North Dakota on the road.

In late January, I suggested that the Bulldogs would need to win 8/9 out of their final 11 to end up in the 14-16 range. In the interim, they’ve delivered a 3-2 performance, so it’s a modest upside surprise that they might be able to afford two more losses.

minnesotaduluth

A few top teams are starting to look like locks for NCAA tournament

Leave a reply

For the first time this season, the top PWR teams are starting to look like locks for the NCAA hockey tournament. Aside from current PWR ranking, the biggest determinant in how possible it is for a top team to fall out of contention is how many games it has remaining (because when looking for the worst case, each is a potential loss). Some teams are down to 4 regular season games remaining while others have as many as 8 (plus up to 2 conference tournament losses, in the worst case, for most teams).

While it’s tough to imagine #1 Quinnipiac or #2 St Cloud St losing out, including a pair in the conference tournament, even that probably wouldn’t be enough to bump them (dependent on how many conferences’ tournament champions come from outside the top of the PWR).

quinnipiac scsu

While #3 North Dakota is mathematically vulnerable, getting swept 4 more weekends (including into the conference tournament) seems unlikely.

und

With only 4 games remaining, #4 Boston College and #5 Providence are near locks (again, dependent on how many conferences’ tournament champions come from outside the top of the PWR).

bc providence

#6 Michigan is the highest ranked team with a significant chance of falling to the bubble, largely by virtue of having more games remaining than any other similarly ranked team.

michigan

Despite a brutal schedule of series against #5 Providence and #11 Boston University, #7 Notre Dame won’t fall much with only 4 regular season games remaining. But, 11-12 going into the conference tournament would leave the Irish vulnerable to being excluded with a poor performance.

notredame

With 6 regular season games remaining for each, it’s too early to call #8 Denver and #9 Nebraska-Omaha safe. It would take a slump for either to miss the NCAA tournament, but they do play each other head-to-head and the Mavericks have a particularly brutal schedule (#8 Denver, #3 North Dakota, and #2 St Cloud).

denver uno

Who’s in position for the NCAAs with four weeks left in the regular season?

Leave a reply

With most conferences having just four weeks of games remaining before their tournaments begin (the Big Ten has five), the field is tightening up a bit compared to my first look at the cutlines.

Still, no one is mathematically a lock — leaving the regular season in the 10-14 range, as is possible for even the top teams, is not safe because each can accumulate two additional losses and no wins in the conference tournament (only in the Big Ten conference tournament is the worst case scenario exiting immediately with a single loss and no wins).

#1 Quinnipiac
#2 St Cloud St
#3 North Dakota
#4 Boston College
#5 Providence
#6 Michigan
#7 Notre Dame
#8 Boston University
#9 Nebraska-Omaha
#10 Yale
#11 Harvard
#12 Denver

Through #12 Denver should be safe for an at-large bid unless they slump and sink below .500 in their remaining games. Teams near the top have more margin for mistakes than near the bottom.

qu

denver

From #13 Mass.-Lowell through #26 Minnesota-Duluth can position themselves for an at-large bid, with those near the bottom requiring near perfect records.

#13 Mass.-Lowell
#14 Cornell
#15 Penn St
#16 Dartmouth
#17 Clarkson
#18 Michigan Tech
#19 Robert Morris
#20 Rensselaer
#21 Minnesota St
#22 Minnesota
#23 Bowling Green
#24 Miami
#25 St. Lawrence
#26 Minnesota-Duluth

masslowell

umd

#27 Northeastern and below would need near perfection and some luck to sneak into position for an at-large bid. Even then, success in the conference tournament would be required to not fall out of position. These teams should plan to do well in their conference tournaments.

#27 Northeastern
#28 Ferris St
#29 Union
#30 Northern Michigan
#31 Air Force
#32 Holy Cross
#33 Bemidji St
#34 New Hampshire
#35 Vermont
#36 Western Michigan
#37 Ohio St
#38 Wisconsin
#39 Mercyhurst
#40 RIT
#41 Merrimack
#42 Bentley
#43 Connecticut
#44 Colgate
#45 Massachusetts
#46 Alaska Anchorage
#47 Michigan St
#48 Maine
#49 Colorado College
#50 Army
#51 Lake Superior
#52 Princeton
#53 Brown
#54 Sacred Heart
#55 Canisius
#56 Alaska
#57 Alabama-Huntsville
#58 Niagara
#59 Arizona St
#60 American International

northeastern

These lines are approximate because it’s entirely possible for a currently lower ranked team to have a better chance of a higher finish than a higher ranked team. Individual teams’ records, remaining games, and opponents can result in different potentials. For example, most of the teams in the 30s have literally no chance of rising onto the bubble, see #35 Vermont, but then you occasionally stumble across a chart like #36 Western Michigan.

vermont

westernmichigan

Methodology

Forecasts include the results of games played through Tuesday of this week.

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

North Dakota eyeing taking #1 from Quinnipiac, but don’t count out St. Cloud

Leave a reply

#1 Quinnipiac has held the top PWR ranking for as long as it has been calculable this season, and the top three have been unchanged for about a month. However, the Bobcats haven’t been perfect and last week’s forecast revealed the first possibility of a new leader. QU held onto its top ranking with a win and a tie, but the odds of someone overtaking them continued to increase.

history

Though Quinnipiac hasn’t lost since December 12, a string of ties (4 out of 7 games in 2016, including 3 of the last 5) has chipped away at their once formidable PWR lead.

On a 5-0-1 streak until last weekend, #2 St. Cloud St. has been nipping at the Bobcats’ heels the last couple weeks. However, St. Cloud’s position also weakened a bit with a split last weekend.

#3 North Dakota, in contrast, has continued to strengthen its position with a 12-1-1 run since Thanksgiving. With St. Cloud St. idle this weekend, UND is the most likely beneficiary of any additional stumbling by Quinnipiac. Not only do the Fighting Hawks stand their first measurable chance of taking #1 this weekend, but it’s a pretty realistic chance.

Even if Quinnipiac wins both its games this weekend, North Dakota can give itself about a 28% chance of taking #1 by also sweeping. A complicated mix of other game outcomes would determine the RPI leader between QU and UND, who would then also win the PWR comparison, and take #1.

If Quinnipiac loses one of its two games this weekend and North Dakota sweeps, the Fighting Hawks are almost certain (over 99% of scenarios) to take #1. A single win doesn’t seem likely to do it for UND, as that would make it difficult to overtake idle St. Cloud St.

If the Bobcats keep up their undefeated streak, but earn another draw, that will also open the door for North Dakota. Those scenarios give UND about a 95% chance of taking #1 with a sweep.

Finally, though they don’t control their own fate, St. Cloud St. still stands a chance of taking #1. If UND loses one, the Huskies would have about a 33% chance of taking #1 with one Quinnipiac loss, or about a 94% chance with two losses by QU (again, those would come down to complicated RPI movements based on other games).

qu scsu und

Weekend outlook – a new #1?, St. Lawrence, and the bubble watch

Leave a reply

Could there be a new #1?

Coming off 3 ties in the last 6 games, Quinnipiac is just damaged enough that falling from #1 has become a realistic possibility. Losses to both Cornell and Colgate this weekend would make a fall to #3-#4 quite likely.

qu

What’s up with St. Lawrence?

After a 1-6 run dropped #27 St. Lawrence from #9 to #26, a pair of wins last weekend failed to advance them in the PWR (on the bottom end of my prediction for that scenario).

StLawrenceHistory

However, you should think of those wins as creating additional ranking potential in the Saints, such that if they can keep winning we could see that potential unleashed in a big jump. Coming out of the weekend between #18 and #22 is reasonably likely if they can notch victories over Rensselaer and Union.

stlawrence

Weekend bubble watch

Looking at #9 Nebraska-Omaha through #18 Minnesota State, only UNO and Dartmouth have really overachieved and underachieved to get there (and Dartmouth wouldn’t be at all surprising if you instead started looking on Jan 4, after a pair of wins vaulted them from #32 to #21).

bubblewatch

#9 Nebraska-Omaha should have a pretty quiet week off, though might slip a little (particularly if Notre Dame experiences success over Vermont and/or BU defeats Mass.).

uno

#10 Notre Dame and #11 Mass.-Lowell face typical outlooks for their positions — the possibility of a small rise with a pair of wins, or a bit steeper fall with a pair of losses.

notredame

masslowell

#12 Denver shouldn’t move much while idle, though might climb a little if Mass.-Lowell and/or Notre Dame falter.

denver

After tumbling a bit from a split last weekend, #13 Yale is unusual in this pack for facing a little more upside than downside potential.

yale

#14 Rensselaer is standing on the edge of a cliff — a couple wins should strengthen its position without much of a corresponding rise in PWR, while a couple of losses could send them plummeting toward 20. Looking at their PWR details, the elevated risk seems to come from a large number of ties in COP that will turn comparisons against them with losses to St. Lawrence and Clarkson.

renssalaer

#15 Dartmouth, #16 Cornell, #17 Penn St, #18 Minnesota St all face noticeably more downside risk from losing than upside potential from winning this week.

dartmouth

cornell

pennstate

mankato

Methodology

Forecasts include the results of games played through Tuesday of this week.

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