2016-17 WCHA Series 31 Predictions

2016-17 Week 10, Series 31: Ferris State (1523) at Alabama-Huntsville (1373)

UAH sweeps7.55%
UAH 5 points (W, OTW)3.16%
UAH 4 points (W, OTL or OTW, OTW)3.25%
FSU Sweeps43.47%
FSU 5 points (W, OTW)8.12%
FSU 4 points (W, OTL or OTW, OTW)7.71%

These two teams met during the first week of the season, and two months later, each team’s prospects have changed radically.  Alabama-Huntsville’s 2016-17 season looks a lot like last season’s so far: a hot October followed by a long slide into the cellar.  Ferris State’s rebound comes with this year’s stellar freshman goalie knocking last year’s out of the crease.

ABOVE likes Ferris to do well in Huntsville this weekend, taking away four or more points 59% of the time and coming away with two or less just 14% of the time.  The Chargers will look to make the New York Times out as prophetic.

Geof’s subjective prediction: Ferris sweeps, denying UAH an A tiebreaker over the Bulldogs.

2016-17 WCHA Series 30 Prediction: Alaska at Minnesota State

2016-17 Week 10, Series 30: Alaska (1464) at Minnesota State (1582)

MSU Sweeps38.98%
UAF sweeps9.86%
MSU 5 points (W, OTW)7.61%
MSU 4 points (W, OTL or OTW, OTW)7.11%
UAF 5 points (W, OTW)3.71%
UAF 4 points (W, OTL or OTW, OTW)3.70%

This is a battle of teams that need to make a statement.  Are the Mavericks still in the top tier in the league?  Is Alaska on the way out of the basement?  BELOW and ABOVE say the former.

Mankato can rout the Nanooks (10.85% probability) to move to 1619 and drop UAF to 1427.  The reverse is a 1.25% probability but would have Alaska jump to 1536 and past the Mavs at 1510.

Geof’s subjective prediction: Mavericks win in regulation, win in overtime.

2016-17 WCHA Series 29 Prediction: Bemidji State at Michigan Tech

2016-17 Week 10, Series 29: Bemidji State (1713) and Michigan Tech (1697)

MTU sweeps36.21%
BSU Sweeps24.46%
MTU 5 points (W, OTW)8.14%
BSU 5 points (W, OTW)5.96%
MTU 4 points (W, OTL or OTW, OTW)5.44%
BSU 4 points (W, OTL or OTW, OTW)2.81%

The above is not a typo.  These are late because I was checking this so much.

Why is Michigan Tech expected to win?  It’s an artifact of how BELOW works.

Two teams play: their matchup creates an interaction.  Win the interaction and you get some level of points.  If the lower-ranked team wins, they have the chance to flip the script and be the better team.  Because these two teams are so close in BELOW, even if not the standings, Tech can flip into the favored team very easily.  Currently, BSU is only picked for a favorable result at a 52% clip on Friday; if MTU can flip that, especially by a big value, there’s a lot of possibility for change.

Things to note:

  1. Bemidji’s embarrassing sweep by Princeton has nothing to do with BELOW.
  2. Michael Bitzer hasn’t been abducted by aliens (aliens don’t interact with Earth as of yet).

Geof’s subjective prediction: Beavers sweep.  (I hate everything.)

2016-17 WCHA Series 28 Prediction: Bowling Green at Lake Superior

ABOVE Prediction for Series 28: Bowling Green at Lake Superior

2016-17 Week 10, Series 28: Bowling Green (1602) at Lake Superior (1457)

BGSU Sweeps42.73%
BGSU 5 points (W, OTW)8.11%
LSSU sweeps8.11%
BGSU 4 points (W, OTL or OTW, OTW)7.61%
LSSU 4 points (W, OTL or OTW, OTW)3.32%
LSSU 5 points (W, OTW)3.19%

As I consider this series, it’s about one thing to me: are these teams who we thought they might be earlier in the season — Lake Superior streaking, Bowling Green barely slipping the surly bonds of earth — or who they are now (BG 3rd in the standings, LSSU in 7th)?  BELOW is convinced that the Falcons have left the nest and thinks that the Lakers are staying close to harbor as the ice edges closer to land.

A great weekend helps the Falcons keep pace in the WCHA standings, probably wrapped up in the hopes that Michigan Tech can do what Minnesota State could not: defeat Bemidji State.  If the Huskies cannot level things at one end of the UP, Bowling Green could take a long stride toward taking the MTU’s place as the top challenger.

BELOW swings: if BG routs the home team in both games (12.30% probability), they jump to 1635 in BELOW as Lake Superior dips to 1424.  If the Lakers rout the Falcons both nights (0.75% probability), they jump to 1533 and pass the Falcons, who would fall to 1526.

Geof’s subjective prediction: Split.  I’m just not 100% sold on this Falcon team’s ability to play at a high level every night, and Gordon Defiel can steal a game.

2016-17 WCHA Series 27 Prediction (ABOVE 1.3)

ABOVE Prediction for Series 27: Northern Michigan at Alaska-Anchorage

2016-17 Week 10, Series 27: Northern Michigan (1397) at Alaska-Anchorage (1312)

NMU Sweeps34.22%
UAA sweeps12.77%
NMU 5 points (W, OTW)7.10%
NMU 4 points (W, OTL or OTW, OTW)6.64%
UAA 5 points (W, OTW)4.32%
UAA 4 points (W, OTL or OTW, OTW)4.26%

After weeks where picking up any points at all were unlikely occurrences (29% against Michigan Tech, 40% against Bowling Green), Alaska-Anchorage has a weekend hosting a struggling Northern Michigan team where a split or better is a 50-50 proposition.  Matt Thomas’s squad has truly struggled this year, but so has Walt Kyle’s.

These two teams are well at the bottom of the WCHA standings, and while it’s early days yet, neither team needs to dig this deep of a hole.  With 29 goals scored in just 18 combined games, this one could be a snooze-fest, especially to Wildcat fans propping their eyes open for the late games.

For BELOW swings: two NMU routs would raise their BELOW to 1438 and drop UAA’s to 1271 split(9.02% probability).  The reverse would bump UAA to 1379 and drop NMU to 1330 (2.04% probability).  As always, there’s a lot of room for movement in both directions when two teams are fairly close in BELOW.

Geof’s purely subjective prediction: UAA wins in regulation and loses in a shootout.

2016-17 Week 9 BELOW Ratings

2016-17 WCHA BELOW Ratings, Week 9

2016-17 BELOW, Week 9

TeamWk9 BELOW+/-Wk8 EWP
Above are the 2016-17 BELOW rankings after Week 9 of WCHA conference play (52 of 140 games played).

Michigan Tech romps in Fairbanks

The Huskies ended up returning from Alaska 10 out of 12 points, a record that helped them stay within striking distance for a matchup against Bemidji State this weekend (who stunningly faltered against a doormat Princeton squad).

A 1713-1697 spread in BELOW means that Bemidji is only expected to win 52% of this time in the matchup, a far cry from a 77% expected winning percentage overall.  Anything is reasonable this weekend; a Bemidji sweep is a true assertion of dominance, while a Tech sweep levels the standings at the top and gives the Huskies a shot at the A tiebreaker.  I know that I’ll be watching Friday’s game (since UAH-Ferris isn’t until Saturday-Sunday).

Bowling Green Slips Against Anchorage

Seawolves fans, this isn’t intended as as slight, but Alaska-Anchorage taking Bowling Green past 65:00 is a negative for the Falcons.  Sure, BGSU-5Pts was the third-highest probability in the table, but when sweeping the league’s worst team was 7.8x more likely, that’s no bueno.

Weeks 10 and 11 both have full slates, and when we round out those two weeks, we’ll be to the halfway point of the WCHA season.

Ahead in ABOVE Predictions

I’ll be doing series-by-series predictions going forward.  It’ll keep the script shorter and allow me to do more runs.  As such, I’ll publish these over the course of the next day or so.

2016-17 WCHA Week 9 Predictions (ABOVE 1.3)

Week 9 Predictions

Short week

Not only is it a short week for teams traveling — especially Alaska-Anchorage, it’s a short week for me after a long one last week.  I spent that week trying to make the code more generic, and so I didn’t get to chain Weeks 9 and 10 together.  Soon, perhaps.

Nothing has changed with the ABOVE model for predicting games.

2016-17 Week 9: Michigan Tech (1676) at Alaska (1485)

MTU Sweeps50.616%
MTU 5 points (W, OTW)6.983%
MTU 4 points (W, OTL or OTW, OTW)6.713%
UAF sweeps5.729%
UAF 5 points (W, OTW)2.282%
UAF 4 points (W, OTL or OTW, OTW)2.265%

Michigan Tech had a tougher-than-expected time in Anchorage last weekend, with only the fourth-most-likely probability (4 points) coming into reality.  The Huskies mushed six hours up the road to Fairbanks to play Alaska, who is a stronger foe.  As such, a 71% of a Tech sweep drops to a 50% one.  That probably matches the expectations Mitch’s Misfits after that overtime loss.

2016-17 Week 9: Alaska-Anchorage (1326) at Bowling Green (1588)

BGSU Sweeps60.435%
BGSU 5 points (W, OTW)7.777%
BGSU 4 points (W, OTL or OTW, OTW)7.753%
UAA sweeps2.495%
UAA 5 points (W, OTW)1.462%
UAA 4 points (W, OTL or OTW, OTW)1.165%

Last week, the Seawolves had about a 2% chance of coming away from the weekend with more points than the visiting team.  This week, they travel to Ohio, where above gives their chances as a bit over 5%.

2016-17 Week 8: Has Bemidji Run Away With This? No, But It’s Close

I rankled some Michigan Tech fans a week or so back when I intimated that they would have a very hard time catching Bemidji State.  I still think that’s true.  Here is my case.

Basic Accounting

  1. Through 12 WCHA games, Bemidji State is 11-1-0-0 for 35 league points.
  2. Through 10 WCHA games, Michigan Tech is 6-1-1-2 for 21 league points.
  3. With 18 games remaining, Tech can finish with a maximum of 75 league points.
  4. Tech and Bemidji still play four times this season; in this scenario, Tech would win the A tiebreaker 4-0-0.
  5. With four losses, Bemidji cannot win enough league games to best Tech; the Beavers would max out at 12-0-0-4 for 73 league points.

Is Tech running the table a ridiculous enterprise?  No, but it is an unlikely one, starting with the fact that they have to beat an unbeaten team four times.

With that said, I want to point out something about point #5 above: Bemidji would be finishing with .750 hockey down the stretch; as of Week 8, BELOW expects BSU to beat any average WCHA team 77% of the time.  So even the elite scenario for the Huskies — winning out, sweeping the Beavers — only just barely gets them past the post.

But you deal in probabilities!

Yes, I do.  Let’s talk about one.

Does BELOW directly deal in things like injuries and suspensions?  No, it doesn’t, but let’s go with this ridiculous example:

  1. Michael Bitzer is abducted by aliens and the entire Bemidji goaltending cadre is left with noodly-armed Muppets.  They’re going to get blown out, right?
  2. Tech is blown out both games in Alaska and finishes with a 1593 BELOW, causing Mitch’s Misfits to show up at my house with flaming pitchforks.
  3. Tech levels the Hoover-goalies with two blowout wins.  Tech BELOW 1666, BSU BELOW 1640.
  4. Lake Superior, which still does have a solid scoring offense, polishes their peashooters and also destroys the Beavers, who are now 11-1-0-4, still with 35 points, and still in 1st place … with a BELOW of 1558.

At 1558, BSU would be behind Tech, Minnesota State, and Bowling Green in BELOW (in some order).  Instead of being expected to win 77% of the time, it would be expected to win just 58% of the time — and that’s after losing any chance at the A tiebreaker with Tech.

Is this scenario likely?  I don’t believe in aliens, but it’s entirely possible that Bitzer could be injured during practice and be out for the season.  It’s possible that his backups could all be terrible, although Reid Mimmack and Jesse Wilkins have a solid track record.  It’s also possible that all three Fitzgeralds could be injured along with any of the seniors leading the Beavers in scoring.

It’s going to take something like that for the Beavers to just fall off the table.  The above scenario requires Tech to get blown out by Fairbanks this weekend, thus lowering their BELOW and maximizing their gains with a sweep of Bemidji.  It’s … far-fetched.

But back to probabilities

ABOVE calculates the probabilities of these sorts of events happening — not abductions in specific, but changes in valuation of the team in general.  BELOW is an Elo-style rating, so all you know from the number is that the system thinks that a team is only so good relative to the average squad (1500).  The probability of a blowout in either direction, however small, is non-zero, and those blowouts do have a big impact.  When the needle moves, the probabilities later in the season can change significantly.

The crazy scenario above has Bemidji losing its next four league games.  BELOW says, on average, that the Beavers would lose four games in their final 16.  But losing the next four games, especially in spectacular fashion, changes BELOW’s opinion of the Beavers significantly.  Once that changes, ABOVE is going to assign different probabilities to the Beavers, and that will be reflected in the results.

Let’s look at the opposite scenario: Tech blows out Fairbanks (Tech BELOW 1704), Bemidji blows out Tech (BSU 1766, Tech 1651), Bemidji blows out Lake (BSU 1783). Instead of being -155 in BELOW, BSU would be +70.  That’s because BELOW already thinks that Bemidji is white hot.

Lather, Rinse, Repeat

A Monte Carlo simulation is the way to figure these things out, and that’s what I use.  I estimate the chances something could happen, I roll a die, I re-calculate, and I roll again.  I’ve only use this technique within a series — simulation 1 says the home team routs on Friday, simulation 2 says the away team wins in a shootout — but it’s not difficult to extend the concept forward.

If we look at the next two weeks, all four teams that play in Week 9 also play in Week 10.  I could estimate the probabilities for the Week 10 games now, but I won’t know what we’ll know after Week 9.  But I can come up with a range of probabilities for the Week 9 games and input them into a simulation of Week 10 … and then forward through the rest of the season.

(If you’re curious, Bemidji would have to keep its undefeated run going through its trip to BG to go above 1800.  It’s harder to move farther from the mean the farther from it you start.)

Has Bemidji sewn this up?  From an account perspective, they haven’t, but it’s getting closer.  Forced to give a guess, I’d say that it’s about 80% likely that Bemidji State has already won the MacNaughton Cup.

I hate everything.

2016-17 Week 8 BELOW Rating

Week 8 BELOW for the 2016-17 WCHA Season

2016-17 BELOW, Week 8

TeamWk8 BELOW+/-Wk8 EWP
Above are the 2016-17 BELOW rankings after Week 8 of WCHA conference play (48 of 140 games played).

Bemidji State Keeps Rolling

The Beavers continued their blitzkrieg through the WCHA, sweeping Northern Michigan despite Michael Bitzer giving up an astonishing three goals on the weekend.   While Michigan Tech has two games in hand on Bemidji, the Green have a 14-point dam built in the standings.

Northern has now tumbled into 8th in BELOW and below 1400, which is not a good place to be.  The Wildcats have just four league points and three wins overall, including two non-conference wins against Wisconsin.  I think that the fans in Marquette are getting restless.

Huskies’ Dynamite Blows Up in their Faces

Simply put, Tech needed to come away from the 49th with six league points to keep pace with the league leader.  Instead, Alaska-Anchorage kept the series close all weekend, including taking the Huskies to a shootout on Saturday, which the Seawolves won.  I try to drive home the concept of probabilities, and while it seemed unlikely, a 4-point weekend for Tech was an 8% probability.  Time and chance happen to them all.

UAA gets on the board with league points this year, although they are well behind their in-state rival with only two games in hand.  The Seawolves have to build on a two-point weekend and believe that holding a top WCHA team to just four goals is a positive sign.

Bowling Green takes care of business

The Falcons look to be back, although plenty of teams have run up the score against Alabama-Huntsville late in the game — although none quite like a six-goal burst in the final 16:15 of regulation.  If there’s a chink in the Falcon’s armor, it’s that UAH’s three goals on Saturday came on the power play.  BG’s PK is in the bottom five nationally through Saturday’s game.

UAH’s 2016-17 looks a lot like their 2015-16: hot start, then falling off the cliff.  UAH has held the lead in the last seven games and has only won twice (their road sweep in Anchorage).  The natives are getting restless.

Ferris State Makes a Statement

Two teams near each other in BELOW that play each other have a big chance for movement: the expected values of the matchups are nearly 50-50, so the bulk of the adjustment factor will be used.  I can explain that in another post, but note that BG got 1/3 less of a bump than Ferris did in their sweep of Lake Superior despite blowing UAH’s doors off.

The Bulldogs made a statement in two different ways: a strong effort in shutting down the Lakers 3-0 on Friday was followed by a wild 7-5 win where half of the goals were scored on special teams and each period had at least three goals scored.  LSSU started Nick Kossoff with Gordon Defiel struggling to date, but the latter picked up the loss despite giving up just one goal.

The standings are more important to the teams than BELOW is (although these posts should totally be on the locker room walls), and Ferris State is now in 5th in league points and just two points from 3rd.  Can the Bulldogs make a return to home ice territory after early-season struggles?

Looking ahead

This week has two league matchups, with Tech finishing the Alaska double with a trip to Fairbanks.  Anchorage comes to Ohio to face BG for a pair for the  weekend’s other slate.  The Huskies (75%) and Falcons (82%) are strongly favored.  Bemidji, Northern, and Lake State all play in non-conference matchups for glory and the hope of raising the WCHA’s woeful PWR hopes.

As far as ABOVE, this time of year is busy for everyone, and I’m no different.  That said, I hope to have a little bit of a surprise for you by Wednesday.

2016-17 WCHA Week 8 Predictions (ABOVE 1.3)

Oh hey, I predict games here.  Doing two jobs in one work week is hard, and this has suffered as a result.  Sorry!  You get what you pay for.

The big change in the model this week is that I’m no longer calculating a center point between each team’s expected winning percentage.  This pushed predictions to the center, which made sense in a tie-happy WCHA.  With just four games going past 65:00 this year, I’m comfortable with not squashing things to the middle.  This will make results a bit more … one-sided.  We’ll see if that was the right call or not soon enough.

The small change in the model this week is that the code’s a little more generalized.  I figure that I can beat on it a little bit more in the next week and have something worth publishing soon.

2016-17 Week 8: Bowling Green (1559) at Alabama-Huntsville (1402)

BGSU sweeps46.088%
UAH Sweeps7.866%
BGSU 5 points (W, OTW)6.768%
BGSU 4 points (W, OTL or OTW, OTW)6.245%
UAH 5 points (W, OTW)2.647%
UAH 4 points (W, OTL or OTW, OTW)2.624%

Yes, even with the Chargers’ strong start and the Falcons’ slow one, BELOW thinks that the visitors are far stronger.  I want the data either way, but the degree that hangs on my wall is not a fan of orange and brown.

2016-17 Week 8: Michigan Tech (1680) at Alaska-Anchorage (1322)

MTU sweeps70.752%
MTU 5 points (W, OTW)8.461%
MTU 4 points (W, OTL or OTW, OTW)8.021%
UAA 5 points (W, OTW)0.771%
UAA 4 points (W, OTL or OTW, OTW)0.759%
UAA Sweeps0.697%

After a week when I really, really rankled Michigan Tech fans by continuing to argue that Bemidji State is making it very, very hard to catch them, I don’t know what to think that they’ll say about ABOVE’s predictions here.  Will it be, “Thank you!  We’re better than you say that we are!” or “You’re jinxing us!”  Either way, ABOVE says that Anchorage will have several trails of paws on it.  Sorry to the Bronies.

2016-17 Week 8: Northern Michigan (1419) at Bemidji State (1697)

BSU Sweeps63.072%
BSU 4 points (W, OTL or OTW, OTW)8.004%
BSU 5 points (W, OTW)7.875%
NMU sweeps1.972%
NMU 5 points (W, OTW)1.275%
NMU 4 points (W, OTL or OTW, OTW)0.898%

I don’t know if the Puckheads with pitchforks are coming for me or Walt Kyle.

2016-17 Week 8: Lake Superior (1502) at Ferris State (1478)

LSSU Sweeps26.782%
LSSU 5 points (W, OTW)4.914%
LSSU 4 points (W, OTL or OTW, OTW)4.607%
FSU sweeps16.592%
FSU 5 points (W, OTW)4.298%
FSU 4 points (W, OTL or OTW, OTW)4.335%

I’m not ready to be regularly publishing this information just yet, but let’s look at something that ABOVE can tell us:

LSSU: Max: 1552 (6.413%) | Min: 1444 (2.947%) | Mean: 1503 | Median: 1501


Max: 1536 (2.947%) | Min: 1428 (6.413%) | Mean: 1476 | Median: 1478

What does that tell you?  Add that to the above table, and you’ll see that it’s a little bit more than twice as likely that the Lakers will blow the Bulldogs out twice than the reverse.  That said, on average, the teams pretty much leave the weekend with the same BELOW that they came in.

As always, these eight games were simulated 100,000 times, but each will only be played once in the next 48+ hours.  Drop the puck!