Predictions for the 2018 college hockey tournament

These predictions are based on the college hockey rankings, which originate from offense and defense rankings for 60 Division I teams. These offense and defense rankings imply a goal rate for an offense against an opposing defense.

Based on these goal rates, I use a Poisson process model to calculate the probability for a win, loss or draw for each game in regulation. If a game is tied, it goes to an overtime in which the first team to score wins. This is not accounted for in these results.

In the results below, 3 seeds Northeastern, Clarkson and Minnesota Duluth are predicted to have a better chance to beat their 2 seed opponents. The tournament starts on Friday, March 23rd, 2018.


Air Force vs St Cloud State: St Cloud State has a 77.6% chance to win. Air Force has a 10.7% chance to win. There is a 11.7% chance for overtime.

Minn-Duluth vs Minnesota State: Minn-Duluth has a 42.7% chance to win. Minnesota State has a 39.3% chance to win. There is a 17.9% chance for overtime.


Princeton vs Ohio State: Ohio State has a 51.5% chance to win. Princeton has a 32.8% chance to win. There is a 15.7% chance for overtime.

Denver vs Penn State: Denver has a 54.3% chance to win. Penn State has a 30.4% chance to win. There is a 15.3% chance for overtime.


Boston University vs Cornell: Cornell has a 53.0% chance to win. Boston University has a 28.9% chance to win. There is a 18.1% chance for overtime.

Northeastern vs Michigan: Northeastern has a 52.0% chance to win. Michigan has a 32.6% chance to win. There is a 15.4% chance for overtime.


Michigan Tech vs Notre Dame: Notre Dame has a 59.1% chance to win. Michigan Tech has a 24.0% chance to win. There is a 16.9% chance for overtime.

Clarkson vs Providence: Clarkson has a 46.5% chance to win. Providence has a 34.3% chance to win. There is a 19.2% chance for overtime.

College hockey rankings prior to the 2018 tournament

These rankings are based college hockey games through March 17, 2018. A least squares algorithm ranks teams on offense and defense based on goals per game adjusted for schedule.

The offense and defense ranks get combined to give the team rankings below. The rating, or the number to the right of each team, gives a predicted average margin of victory against an average Division I hockey team.

Predictions based on these numbers will be posted soon.

Please leave your complaints in the comments.

1. St Cloud State, 1.88
2. Denver, 1.70
3. Minn-Duluth, 1.53
4. Cornell, 1.49
5. Minnesota State, 1.42
6. Clarkson, 1.31
7. Ohio State, 1.28
8. Northeastern, 1.25
9. Notre Dame, 1.10
10. North Dakota, 1.03
11. Providence, 0.99
12. Boston College, 0.94
13. Penn State, 0.88
14. Boston University, 0.80
15. Minnesota, 0.77
16. Princeton, 0.70
17. Harvard, 0.68
18. Michigan, 0.63
19. Union, 0.48
20. Omaha, 0.33
21. Western Michigan, 0.27
22. Wisconsin, 0.26
23. Maine, 0.25
24. Miami, 0.23
25. Quinnipiac, 0.16
26. Connecticut, 0.16
27. Mass-Lowell, 0.11
28. Michigan Tech, 0.09
29. Bowling Green, 0.06
30. Northern Michigan, -0.01
31. Colgate, -0.08
32. Yale, -0.09
33. Bemidji State, -0.12
34. Merrimack, -0.14
35. Massachusetts, -0.16
36. Colorado College, -0.17
37. Michigan State, -0.23
38. Air Force, -0.30
39. New Hampshire, -0.34
40. Dartmouth, -0.44
41. Mercyhurst, -0.46
42. Canisius, -0.47
43. Robert Morris, -0.64
44. Vermont, -0.65
45. Army, -0.66
46. Holy Cross, -0.91
47. St Lawrence, -0.95
48. Alaska, -0.96
49. RIT, -0.98
50. Brown, -1.06
51. Bentley, -1.08
52. Ala-Huntsville, -1.15
53. Sacred Heart, -1.24
54. Ferris State, -1.26
55. RPI, -1.26
56. American, -1.29
57. Arizona State, -1.31
58. Niagara, -1.38
59. Lake Superior, -1.59
60. Ala-Anchorage, -2.08

Is the RPI or preseason polls a better predictor of NCAA tournament success?

Data is king. In 2018, our tech filled world demands collecting data to make meaningful predictions.

The RPI, the Rating Percentage Index used by the tournament selection committee, does use data to rank 351 college basketball teams.

The preseason polls do not. Pollsters submit their ballots before seeing any games much less collecting data on them.

So the RPI must be a better predictor of tournament success than the preseason polls, right?

Over on FiveThirtyEight, I wrote about how the preseason polls are better predictors of tournament success than the RPI. The wisdom of crowds beats bad mathematics.

In the article, I highlighted three teams from the current season.

First, I noted how Duke and Michigan started the season ranked 1st and 2nd respectively in the preseason AP and Coaches polls. This suggests both of these teams have the talent to make a deep tournament run despite not living up to their potential during the regular season.

Duke looked fantastic this past weekend, beating a solid Rhode Island team 87-62 to make the Sweet 16. In contrast, Michigan State shot 26% from the field to get stunned by Syracuse.

The article also highlighted Virginia, a team the committee deemed the number one overall seed but didn’t make the top 25 in either preseason poll. Before the tournament started, I had questioned whether Virginia had the NBA level talent necessary to win the tournament.

In no way does this suggest Virginia would lose to 16 seed UMBC. However, the preseason poll suggested this was a tiny bit more likely than any of us expected.

To read the article on FiveThirtyEight, click here.

Podcast: Adam Stanco on predicting the NCAA tournament champion, Part 2

I continue my conversation with Adam Stanco, producer at the Pac-12 Network and hoops expert, on teams that can win the 2018 NCAA tournament.

In Part 1, we discussed Virginia, Villanova, Duke and Cincinnati. In this Part 2, we get into these topics:

  • The player that needs to excel for Michigan State to make a deep run
  • The reason why North Carolina could be better than the numbers indicate
  • The problem Kansas has on defense
  • How Michigan makes a deep run in this tournament
  • Two other teams that Adam finds intriguing as championship contenders

To listen on iTunes, click here.

To listen here, click on the right pointing triangle.

Podcast: Adam Stanco on predicting the NCAA tournament champion, Part 1

On this episode of the podcast, Adam Stanco, producer on the Pac-12 Network and hoops expert, joins me to break down the teams that can win the 2018 NCAA tournament.

We focus on primary contenders because the choice of champion is the most important choice you will make in filling out your bracket.

Among other topics, we discuss:

  • The flaw that might hamper Villanova
  • How much NBA talent does Virginia have on its team?
  • The reason why Duke’s zone defense might work
  • How Duke’s defense rank has changed since going to the zone, and why it might be a fluke
  • Can Cincinnati score enough to win 6 games?

This is part 1 of my conversation with Adam. In Part 2, we will break down another set of teams that could win the tournament.

To listen on iTunes, click here.

To listen on the site, click on the right pointing triangle.