Nicole: In case you missed it in the first Wednesday Women of the year, the format is a little different this year. I’ll be joined by guest writers throughout the season to share their unique insight. This week I’m joined by Grant Salzano, Boston College alum and writer at BC Interruption whose Pairwise Predictor is an indispensable tool as the season winds down.
Grant: Glad to be here and glad some folks find the calculator useful! I know the annual changes from the NCAA make it tough to keep up with, but it’s been nice to be able to see how the committee is ranking the teams as the season goes on.
Nicole: You put out three different calculators. Can you explain what each one does, why it’s worthwhile to look at all three and what fans should take away from glancing at those calculations week in and week out?
Grant: Sure, so the three calculators are KRACH, the Pairwise, and a ranking of my own called the “GRaNT” Rankings (which stands for Grant’s Reasonable and Not Terrible Rankings and is only slightly tongue-in-cheek). First and foremost, let me just make it perfectly clear that as someone with a firm grasp on how these formulas are calculated, I find KRACH to be far and away the most sound ranking system. It’s a little difficult to explain with an “Explain Like I’m 5” description, but essentially what it’s doing is setting each team a rating so that when you add up how many games a team with that rating should have won, it exactly equals the number of games they actually won. It’s very satisfying mathematically and does a really good job of accounting for schedule strength.
One thing I like about KRACH is that it makes it very clear that anyone can beat anyone on a given day. I’m (obviously) a big numbers guy, and as such I sometimes have to remind myself that these games are played by people, and not even professionals, either, who have maximized every bit of their talent and have less variation in how well they play. What I mean is, it’s not as simple as looking at the KRACH rankings, no matter how big the sample size is, and seeing Wisconsin is ranked above Boston College and saying that, therefore, Wisconsin is better than Boston College. In fact, KRACH says that not only can the Eagles beat the Badgers, but that we should actually expect it to happen once in about five tries. It’s a little thing, but it does give a little bit of a glimpse back into the human side of the sport, in a way, despite it being a mathematical formula.
The Pairwise calculator has more practical use just because it mirrors how the NCAA selects the national tournament, and as such is the only one that “matters.” In women’s hockey, the Pairwise is essentially just “RPI NPI with slight adjustments.” There are limitations to the Pairwise’s ability to account for schedule strength, and the adjustments it makes are a bit arbitrary (it is weighted 70% for schedule strength, which, as far as I can tell, is a number that was just pulled from a hat), but it’s what the NCAA uses, so it’s what we need to be following throughout the year.
GRaNT is the same idea mathematically as KRACH but where it looks at goal differential instead of wins and losses. I won’t go into the mathematical details, but you can go into the weeds here if you are a nerd for numbers like I am. Despite GRaNT being my own ranking system, I do much prefer KRACH since the only real goal for a hockey team is to win, not to blow out its opponent.
But I think GRaNT is useful for two reasons – one, just because more rankings is always better (which is why I lament the loss of Rutter and WCHODR in the pantheon of women’s hockey rankings), just like more political polling gives you a clearer view of upcoming elections because it can help smooth out any biases. But also because I think there is some use in seeing that a team is always blowing out or shutting out their opponents. That seems especially true in women’s hockey, at least to me, where you can often get several teams with very few losses and the magnitude of their victories can give you some added insight as to who has been the stronger team.
Nicole: That’s actually all super helpful, even for me. I’m a stereotypical “no math, ever” writer, so having you break it down that way helps me more easily remember and wrap my head around it and I think will help fans see what those different rankings are telling us. Thank you for that.
You spend a lot of time compiling data and crunching numbers between your calculators and compiling votes for SB Nation’s weekly poll at The Ice Garden. Is there anything that has surprised you or stood out? Where do you think the biggest gap exists between public perception of a team and where they actually fall statistically in your calculations?
Grant: So far the biggest surprise has to be the early strength of the top ECAC teams, right? Maybe it’s just that my hopes have gotten beaten down so many times through the years that I’ve finally started to believe the western hype, but it had really seemed to me for the last couple seasons that the gap between the WCHA and everyone else was just getting blown wide open. Now we’ve got Quinnipiac starting off really strong, Yale coming in and knocking them off with a win immediately, Colgate with a hot start, Cornell with a hot start – I mean, these four teams are a combined 30-3-0. Now, they’re only playing peer programs from the east so far… but they’ve done just about all you can ask of them.
To answer your question about the gap between public perception and the rankings, 90% of that answer is going to just come down to sample size. Yale isn’t going to end the season #1 in KRACH, but that’s where they sit just four games in. I suspect Wisconsin is a fair bit better than their 8th place KRACH ranking as well.
But beyond just the obvious “it’s still early,” answer… look, the regional differences are real. Nobody hates saying that more than me. The problem is, even for a ranking system like KRACH that does a decent job of accounting for schedule strength with limited inputs, we have so few games between the top eastern and western teams that the math ends up with a lot of noise trying to correct for it. It’s like if you try to boost the volume of a quiet audio source; you’ll get a lot of distortion in there.
There’s a lot of reconciling someone filling out a ballot – be it for USCHO, USA Today, or SB Nation – needs to do between conferences. Aside from one week where I brought Quinnipiac up to #3, six of my seven ballots for the SB Nation poll have had WCHA teams one through four. From there it’s pretty much all ECAC teams, with a question of where to throw Northeastern in the middle of them. I think the conference stratification, and more specifically the regional stratification between the WCHA and everyone else, is that stark, and to me it’ll take a lot more than just one month of games to help turn that trend, especially when it’s ECAC teams are the ones beating ECAC teams and WCHA teams are the ones beating WCHA teams.
Nicole: Let’s talk about Hockey East.
Grant: That’s very mean of you.
Nicole: You do this to yourself.
I definitely want to talk about Vermont, Providence and Northeastern, but let’s start with your Eagles.
Grant: Wow that’s even more mean.
Nicole: Like this is my fault?
I know you had high hopes for Boston College heading into the season based on how good the roster looked on paper. That has not really panned out. What have you seen from your favorite team that accounts for some of that dissonance?
Grant: I hate to keep bringing up the east/west stuff but I think there’s a part of that here as well. I don’t follow youth/high school/club hockey all that closely, but from what I can tell, a lot of the U17 and U18 players are coming more and more from out west. Then of course those girls get to have more top-end competition which makes them even better. Boston College has 16 Massachusetts natives out of the 26 players on the roster. Look at the programs that have surpassed BC over the last few years: Northeastern has 6 players from Massachusetts. Colgate has 4 players from all of New England + New York combined. Quinnipiac has just 4 as well.
What that tells me is that as the sport has grown, and the talent pool has expanded both nationwide and internationally, Boston College trying to lock down Massachusetts recruiting has seen us fall behind. When I see BC “look good on paper” it’s a result of a lot of our local recruits looking good against local competition. And that’s really not a knock on those players or even a knock on Massachusetts youth hockey – it’s just that there is now so much more talent everywhere else that Massachusetts isn’t the hotbed it used to be.
Who are BC’s top players? Hannah Bilka, Texas. Abby Newhook, Newfoundland. Cayla Barnes, California. Abby Levy, New York, but played at Shattuck in Minnesota. You’ve got one truly elite player in Kelly Browne who’s from Massachusetts, and when it’s just one despite being where 62% of your roster is from, you can see how the numbers start to fall apart.
And it’s more than just east/west. Northeastern has succeeded largely through tapping into the rapidly growing glut of girls’ hockey talent worldwide. That requires significant investment of resources to get those players to come consistently, and as Boston College does not give need-based financial aid to international students, I can’t imagine it’s worth the time necessary for the coaching staff to search out those international gems (aside from the handful of elite Canadians we’ve brought in) when you’re only talking maybe one player every couple years since they’re going to pretty much have to be on full scholarship unless they pay their own way.
Nicole: I don’t think I knew that about BC and international students. That is actually a really big deal with the changing landscape of women’s college hockey and how teams have been able to elevate themselves when they’re not able to recruit the top American talent. We’ve seen a lot of that in Hockey East – at Vermont and Maine in particular.
Switching gears, I wanted to talk about this both because it’s Hockey East, but also because of your work on the numbers. Northeastern’s fanbase was upset when yesterday’s poll came out and the Huskies were eighth. However, all of your various ranking systems have them seventh and with a pretty big gap, numbers-wise, between them and sixth place. What do you think the disconnect is there?
Grant: As someone who roots for the team that was The Good Hockey East Program Who Hasn’t Won The Natty before Northeastern was, I just know they have a massive chip on their shoulder and feel like they should get more respect than they’re getting. But the pollsters early on have had to consider the fact that Northeastern, as good as they are, have played like a great team that has lost several players in the offseason (including a Patty Kaz winner and another Kaz finalist), and frankly, their results have not been of a team that wants to claim to be a top 4 program.
Who could you rank them ahead of right now? Certainly not Colgate or Quinnipiac out east. And the rest comes from acting like losing to Maine and tying Providence is the same as the top WCHA teams trading losses. It just isn’t, and beating up on UNH is not going to impress anyone, sorry.
Nicole: To piggyback off that point, one issue they are going to deal with is that while the strength of schedule of the teams around them is going to continue to improve theirs is mostly going to remain where it is and they won’t be able to get much in the way of quality win bonus points.
Grant: That’s right. Northeastern has just one game left against a top ten team the entire rest of the season, a game against Cornell in Nashville. That’s it. Just from a numbers perspective, Northeastern would have to nearly run the table the rest of the way just to keep pace with a team like Quinnipiac who can lose some games against Cornell, Colgate, Clarkson, and Yale without losing much ground. That goes for the Wisconsins of the world who get games with Duluth, Minnesota, and Ohio State to bulk up their resume. Northeastern does not even have an opportunity to build a resume.
And one last major problem NU has now – according to my back of the envelope math (and forgive me if my numbers are wrong, I’m having to do this all by hand), the WCHA has a 0.806 winning percentage out of conference, the ECAC 0.725, and Hockey East an incredibly pedestrian 0.529. Probably something like 90% of the non-conference games that will be played this year have already been played, which means these numbers are pretty well baked in at this point and these non-conference records are how the various ranking systems sort through teams from different leagues. The long and the short of it is that Northeastern winning most of its games against teams in a league that barely went 0.500 out of conference is not going to impress anyone.
Nicole: What do you think we’ll see within Hockey East this season? Is Northeastern still a shoo-in to win it all? Are they the only team that will get into the NCAA tournament? How do you think Providence and Vermont stack up?
Grant: I’ll probably have a better answer to that after this weekend when the Eagles play Northeastern. I’d like to have a better idea of how good the Huskies are. My early thoughts are that they are still on their own level relative to the rest of the league, but that they’re more vulnerable to a single-elimination bad game than last year and seeing someone like a Vermont or a Providence or a BC winning in a one-off wouldn’t be the stunner it would have been a year ago.
Despite the expanded national bracket it’s looking like Hockey East might need to have an upset winning in the league tournament to get a second team in – right now both the CHA and NEWHA frontrunners are below the cut line, so a second Hockey East team would have to be 9th in the Pairwise at worst to make it. There’s a long way to go, of course, and it’s possible just from a numbers perspective that someone could emerge as a strong second place team and eke their way in as the top ECAC and WCHA start beating up on each other with the conference schedules starting up in earnest… but I mean, hell, even Northeastern is barely in at #7 right now.
Nicole: We’re now two or three weeks into the Ivy League season and those teams have not disappointed, both in their games and with bringing the chaos. I said this on Twitter, but I’ll go ahead and out myself as someone who said I thought Cornell was going to have a good offense, but that it might take awhile. The Big Red put 15 goals on the board against St. Lawrence and Clarkson, so it’s safe to say I was definitely too cautious in that assessment. Both of their weekend games were surprises in their own right. I definitely didn’t see them giving up six goals – that’s definitely something to keep an eye on.
But the bigger shock was the Big Red hanging an eight-spot on Clarkson. The Golden Knights had some big injuries, but were still handling themselves pretty well. That is a heck of a statement win for Cornell and has to be a bit of a mind-blower for the Golden Knights. I assume they’ll try and shake that off, but it feels like they were super exposed and have some reassessing to do.
I’m trying to hold off on reading too much into Yale’s first few games since it’s such a small sample. They started a week later than the rest of the Ivies, so it’s hard to get a grasp on them overall with just four games to assess. It’s certainly gone well so far. I was worried for them against Quinnipiac with so little time on ice before the game, but they scored four goals on 20 shots and kept the Bobcats in check offensively.
Now that we’ve seen the top teams play each other a bit and everyone but Yale suffered a loss, how are you feeling about how the conference will shake out? Which one of these teams is the most likely to take their league crown and go on to dethrone the WCHA?
Grant: Based on how Quinnipiac played against BC I’m pretty sure they could take down the Bruins on a good day! I laugh so I don’t cry, honestly… haha
In all seriousness, Quinnipiac does seem to have themselves quite a team this year. They returned a whole lot of their scoring and have started out red-hot. Colgate has consistently been knocking on the door, so I’d expect them to be one of the main contenders with the Bobcats. The other teams have more question marks – how does Cornell get going? Can Yale really match their wild out-of-nowhere run from last year? Is Clarkson legit or a paper tiger just beating up on bad teams? Are Harvard and Princeton just having a tough start or are they the top ten teams some folks felt they were in preseason ballots?
Geez, when you actually write it out, it’s pretty wild how deep that conference is.
Nicole: Well I wasn’t planning to talk WCHA, but I feel like we should with St. Cloud handed Minnesota their first loss of the season on Monday. The game doesn’t count for conference standings – it was the Hockey Hall of Fame game, which the Gophers play in every year, while the opponent rotates. Last night’s game took place at Andover High School, a perpetual Minnesota girls high school hockey contender.
This wasn’t a win the Huskies eked out – they jumped out to a 3-0 win, held off three power plays in the second, made 22 blocks and added an insurance goal in the third. Sanni Ahola was absolutely stellar in net, but don’t write this off as a win on the back of the goalie standing on her head. It was a complete effort from St. Cloud.
It was SCSU’s first win over Minnesota since 2010 and their first win over a #1 since 2007. The Gophers had a 63 game unbeaten streak against the Huskies going into the game.
St. Cloud had a tough start to the season, facing Ohio State, Wisconsin and Minnesota three weeks in a row. Since then, they’ve won five straight.
The loss is probably going to light a fire under the Gophers, which doesn’t bode well for their opponent this weekend, Wisconsin. But perception wise, Minnesota’s loss to St. Cloud is actually pretty helpful for the Badgers. UW was pushed to OT by the Huskies a few weeks ago and that had folks raising their eyebrows. The Badgers lost to Penn State in the opening game of the season. The Nittany Lions were ranked, but it was an unexpected loss for Wisconsin. St. Cloud’s win here takes over the top spot in “most unexpected upset.”
Grant: Most unexpected so far anyway. You just know we’ve got about one a week coming this year…
That one result certainly did affect the Pairwise, though you would expect that this early in the season. The Gophers dropped from 2nd (behind the just-started Yale Bulldogs) to 5th. It just goes to show how deep it is at the top this season, and how much parity women’s hockey has gained over the last several years. It was definitely the right time to expand the tournament field.
Nicole: Look at you dropping that nugget right there at the end. I like it.
As a chaos gremlin, I’m fully on board with tacking “yet” onto the end of “most unexpected loss.” This season has featured so many great games already and we’ve got several months to go. The possibilities are endless and it’s going to be so much fun to watch!
