Unlike college football, college hockey manages to have a simple, fair, effective method for selecting their NCAA tournament that takes subjectivity out of the process and treats everyone fairly – and generally yields very few complaints.
Whether you are a big brand like BC or BU, an upstart like Stonehill, or a team that hasn’t won a tournament game since the Reagan administration like Northeastern, everyone is treated the same and knows exactly what they have to do to make the tournament.
While nobody considers college hockey’s Pairwise rankings perfect, it’s hard to argue that it’s not a better system than what we get from most college sports, made particularly acute when we see the giant endless pointless arguments over the last spots, and the common bias toward the biggest brands in sports like football and basketball (though surprisingly SMU did not get screwed over in this instance).
The Pairwise may not be perfect, but college hockey’s system has some great things going for it:
- Everyone has the same exact rules applied to them: if you win your conference championship, you’re in; if you don’t, the math of how your games will impact your chance is the same for every team regardless of whether you are Michigan or Alaska-Anchorage
- Strength of schedule matters massively (more on this later) but winning your games matters the most
- It’s not just a set it and forget it formula – tweaks are made when coaches/programs determine they are necessary. A bonus was added for road victories, to encourage scheduling tough road games; a quality win bonus was added to the RPI calculation to further incentivize strength of schedule; OT wins were treated as 2/3rds of a win instead of a full win; and some criteria have been changed or removed over the years, such as taking out criteria focused on your record in your last 16 games, to instead take the full season into account equally
The other great thing about having a fair, transparent mathematical standard for tournament selection is that it makes the end of the season much more exciting, because you know exactly what the stakes are and what teams need to do to clinch tournament spots, clinch seeds, move up or down, or be eliminated – creating wild ‘game 162’ style moments that are impossible to recreate with a subjective selection process.
We all remember ESPN yammering on about how Georgia ‘would clinch a spot in the college football playoff’ if they beat Georgia Tech a few weeks ago… but they wouldn’t really?, and they might still have been in even if they lost. Wouldn’t it have been more exciting to know for sure? And to have other teams watching knowing that their fate was based on the outcome of that one?
A Pairwise-style system would be great for college football, and I’ll add to my advocacy for it by presenting what the 2024-25 CFP would look like if they used college hockey’s RPI (the main factor in the Pairwise rankings) to select the tournament. Friend of BCI Adam Marchany created a calculation of college football RPI rankings based on hockey’s system.
A few notes on this system:
- RPI is calculated by a combination of your record and your strength of schedule (opponents’ winning percentage + opponents’ opponents winning percentage). We kept this the same for college football.
- The college hockey RPI calculation adds in a quality win bonus for wins over top 20 teams – 5 points for a win over #1, then descending down to 0.25 points for a win over #20. For college football purposes we counted quality win bonus as being against the top 25, for consistency with how teams are rated in the college football rankings.
- There is also a bonus for winning on the road – essentially, a road win is 1.2 times the value of a home win
- We filtered games against FCS teams out of the rankings, which a) seems fair enough and b) somewhat echoes hockey’s method of filtering out wins that would actually reduce a team’s RPI due to how low-quality the opponent is. A tweak would need to be made to take into account losses to FCS teams, but none of the relevant programs this season lost to an FCS team. My suggestion would be a ‘low quality loss penalty,’ a subtraction from raw RPI for losing to a team outside of the P4 or G6.
- We also elected to not count conference championship games in the calculation, since winning your conference title clinches you in anyway, and everyone has generally agreed (including the committee) that teams shouldn’t be punished for losing conference finals
With all that said, here’s what the adjusted RPI rankings look like, starting with the 5 autobids:
Team – RPI ranking – Quality win bonus
1. Oregon (B1G Autobid) – .6604 – .0072
5. Georgia (SEC Autobid) – .6190 – .0092
9. Arizona State (B12 Autobid) – .6066 – .0021
13. Boise State (MWC Autobid) – .5977 – .0000
20. Clemson (ACC Autobid) – .5792 – .0000
–
And your 7 at-large teams:
2. Texas (SEC) – .6210 – .0000
3. Notre Dame (IND) – .6206 – .0008
4. Indiana (B1G) – .6192 – .0000
6. Penn State (B1G) – .6108 – .0005
7. Miami (ACC) – .6091 – .0000
8. SMU (ACC) – .6089 – .0000
10. Ohio State (B1G) – .6013 – .0074
This would give you a bracket of:
5) Texas
12) Clemson
Winner plays #4 Boise State
8) Penn State
9) Miami
Winner plays #1 Oregon
6) Notre Dame
11) Ohio State
Winner plays #3 Arizona State
7) Indiana
10) SMU
Winner plays #2 Georgia
Notably, this would be a pretty similar field to the real one, swapping out Tennessee for Miami but otherwise remaining the same.
Your first teams out:
11. BYU – .6004 – .0039
12. Tennessee – .6000 – .0020
14. Alabama – .5957 – .0068
15. Iowa State – .5955 – .0000
16. South Carolina – .5867 – .0024
Also worth noting that Army (.5743 RPI, .0000 QWB) would have just barely been edged out by Clemson (.5792) for the last autobid, which is kind of hilarious, but also – doesn’t that seem pretty fair? It was a fairly weak Clemson team, with three losses, and a remarkably strong Army team, with 1 loss to a high quality opponent – and Clemson still has the edge, but just slightly. There should be at least some way for a team like Army to get in, and under this system, they 100% would have if they had beaten Notre Dame, or perhaps had one stronger OOC opponent on their schedule who they beat.
Things change somewhat if you include the conference championship games in the final calculation. Your autobids would be:
- Oregon
- Georgia
- Arizona State
- Boise
- Clemson
And your 7 at large would be:
- Notre Dame
- Indiana
- Miami
- Texas
- Ohio State
- Tennessee
- BYU
Tennessee and BYU would be in, and dropping out would be SMU and Penn State after getting dinged by their losses in conference championship games.
I think there’s merit to including or not including conference championship game results, but I’d probably stick to the one that doesn’t count them, for fairness’s sake. You could theoretically create a method that adds to your RPI if you win your conference title game, for seeding purposes, but filters out a loss, to not ‘disincentivize’ making the conference final. Plenty of ways to make it fair, transparent and consistent, rather than up to someone’s subjective judgement.
Either way, this system would create an honest, exciting tournament selection process; and would strongly incentivize scheduling tough games, especially on the road. But more importantly, would incentivize winning. If you’re in a stronger conference, and win your games, you will get in, no matter what – but there’s still a chance even for the Armys of the world. Isn’t that what most people want, outside of SEC and B1G homers?
BC hockey is in a commanding spot in the Pairwise right now, because they scheduled four difficult road non-conference games at Michigan State (x2) and St Cloud State (x2), and won 3 of them. Of course, they could have scheduled 4 easy games at home instead and maybe won all four, but they were further incentivized to schedule those difficult games, which ends up helping them a lot more. But you have to win them and earn that ranking on the ice. It’s a good system, and incentivizes exciting matchups.
That’s the beauty of the college hockey format, and while I doubt anything will ever be put into place that would take out the ability of big brands, TV, etc to lobby for more playoff spots for their cash cows, it would make college football a better sport to put such a system in place.