Your Annual Reminder

In seeing how the U SPORTS women's hockey conferences are playing out, I feel like I need to drop another PSA in here about how the U SPORTS rankings that are published all season mean absolutely nothing. U SPORTS will try and convince you that they're an accurate way of comparing teams across conferences, but I'm here to tell you that's a fallacy because these teams never play one another outside of exhibition and National Championship games. If they don't play one another in meaningful games more often, comparing them is an absolute waste of time other than giving those schools something they can post to social media. Let me say it again: the rankings mean nothing, and they should not be a factor for seedings at Nationals.

The final Top Ten Women's Hockey Rankings using the significantly-flawed ELO system where the points don't matter or make sense showed the following rankings below. The comments in red are how each of them finished their season within their respective conferences. Pay close attention to the comments. They say a lot.

Based upon how everything has finished thus far in U SPORTS women's hockey, just two of the top-ten teams in the nation as per the February 18, 2025 rankings will have won their final conference game this season. Those two teams are Alberta, and one of Waterloo and Toronto based on their final tomorrow. Every other nationally-ranked team will have lost their final game and series this season.

You can make the case that there are teams from the same conference who beat one another and might be closely-ranked to one another such as UBC and Alberta - there was 133 Elo points between them - but things get fun when you consider that StFX was ranked 12th on February 18 before they defeated eighth-ranked St. Thomas and fifth-ranked UNB en route to winning the AUS championship. StFX was 36 Elo points back of St. Thomas and 82 Elo points back of UNB, so the probability modeling that Elo's flawed system suggested is that StFX should have lost to either of these teams in those series.

That didn't happen, though, as StFX needed five games to defeat St. Thomas before sweeping the UNB Reds in the AUS Final. In short, St. Thomas gave StFX everything they could handle before finally falling while higher-ranked UNB dropped a pair of 2-1 games in the final as they were swept out of the playoffs. As stated in the linked articles, Elo doesn't factor in things like hot and cold streaks or player performances, so should we be surprised that it got the probability for StFX winning wrong, especially when they had a better conference record than St. Thomas did this season?

Let's take this one step further as the RSEQ conference final was a masterclass in Elo's flawed predictions. Bishop's played Montreal in a series where second-place Bishop's had a 9-11 record this season while third-place Montreal sported an 8-12 record this season. Clearly, this series appears to be close, and Elo had the two teams separated by 33 points in the rankings. Bishop's swept the semifinal series by 3-2 and 4-3 scores, and they advanced to face second-ranked-team-in-the-nation Concordia. These two teams were separated by 325 Elo points - the largest margin between any conference finalists this season - so Elo was suggesting an easy series win for the Stingers based on the algorithm calculations.

Concordia seemed to prove Elo correct with a 5-4 in Game One, but Bishop's did what Elo doesn't - they made adjustments to their game. The results were 4-1 and 6-3 wins in Games Two and Three, and Bishop's earned their first RSEQ banner and their first trip to Nationals. Concordia gave up 31 goals in the regular season and two more against McGill before Bishop's hung 14 goals in three games on the Stingers. I'm not saying Bishop's will win the 2025 U SPORTS National Championship, but I'd certainly have a lot more respect for them than the Elo system does. They're really good.

You might be saying, "Teebz, that data is a month old," and you'd be correct. Rankings are stopped during playoffs because it makes no sense to rank a team in the Top Ten only to have them lose five days later. This is the problem, though, as illustrated above: teams in conferences play each other a lot and can make adjustments as the season progresses. All of Alberta, StFX, and Bishop's - ranked lower than the teams they faced in the final - prevailed in the final because they made adjustments based on who they were facing.

That familiarity helps to level the playing field when it comes to winning probabilities, and Elo does not account for this. The sample sizes of games between Alberta and UBC is that so much greater than, say, UBC and Toronto, and it helps to explain why Alberta is 5-4 in the last five years of playoff games against UBC. Elo, though, wants you to believe there's a 133-point gap between UBC and Alberta - bigger than the gap between StFX and UNB, and bigger than the gap between Guelph and Toronto (85 points) after Toronto eliminated the OUA's top team. Elo is wrong more than it's right.

Based on what we know and assuming that Waterloo defeats Toronto in the McCaw Cup Final, this is how the 2025 U SPORTS National Women's Hockey Championship rankings could look next week:

Nothing is set in stone yet as the OUA's McCaw Cup Final goes tomorrow at 3pm CT. You can watch it for free by registering for an OUA account on OUA.tv, and we'll know which of the two Ontario teams will occupy the OUA top-four seeding at Nationals. With Waterlook hosting the McCaw Cup Final and the U SPORTS National Championship, I'd expect a fantastic crowd for Saturday's final.

Having three lower-ranked teams winning championships was something not a lot of people would have predicted. That especially goes for the algorithms run by the Elo system for ranking teams in Canada who would have never predicted this based on winning probabilities. And while it's fun to be a ranked team throughout the season, the only rankings that matter are the ones that are earned at the U SPORTS National Championship so let's abolish the Elo system.

The national rankings do not matter. Ever.

Until next time, keep your sticks on the ice!