One Weird Trick Will Help You Predict The NCAA Hockey Tournament

The NCAA hockey tournament is notoriously unpredictable. The day RIT knocked off #1 Denver in 2010, then won again the next day to advance to the Frozen Four, was the day we could pretty much no longer call anything that happens in this tournament an "upset."

Two of the last three national champions were the very lowest ranked at large teams to make the field. The #1 overall seed has only won the national title once in the last seven years.

So what is the best, most tried and true way to pick winners in the NCAA regionals? Pick who's coming in hot based on conference tournament results?

Well, not really. In 2013, Wisconsin was on a Northeastern-like run, storming to the WCHA title and coming in to Manchester as a sexy pick to win the regional and go all the way. They proceeded to get blown out of the water by UMass Lowell in the first round. Last year, Miami was on a similar run, clinching a 1 seed by virtue of their run to the NCHC championship. Providence, who lost in the Hockey East quarterfinals, smoked 'em. 3 of the last 5 national champions did not win their conference tournament.

One factor, however, proves to be pretty predictive time and time again of who will win in the regionals: travel distance. We examined this phenomenon last year, breaking down results for each seed since 2008 depending on whether they were playing in-state, taking a bus out of state (according to NCAA rules, this means a trip less than 350 miles), or if they flew to their regional destination.

Last year furthered this trend. Teams playing in state (hello, Providence and North Dakota) went 4-0. Teams taking a bus out of state went 3-3, including #1 seed BU who won their regional. And teams who had to fly went 5-9.

Overall since 2008:

Teams playing in-state are 29-16 (.644)
Teams traveling out of state by bus are 27-20 (.574)
Teams flying are 40-60 (.400)

1 seeds are 11-1 in state, 17-3 by bus, 7-13 by flight, 35-17 overall (.673)
2 seeds are 9-9 in state, 1-3 by bus, 11-13 by flight, 21-25 overall (.457)
3 seeds are 4-4 in state, 9-10 by bus, 6-15 by flight, 19-29 overall (.396)
4 seeds are 5-2 in state, 0-4 by bus, 16-19 by flight, 21-25 overall (.457)

The big takeaway from these numbers is that if a #1 seed is playing close to home, you should pick them to win their regional. #1 seeds either playing in their home state or nearby are dominant in the tournament, with a 28-4 record since 2008.

Once you put 1 seeds on a plane, they don't fare any better than anyone else, with a winning percentage of .350 -- actually worse than the winning percentage of #2 and #4 seeds that fly to their regionals.

If you factor out 1 seeds playing close to home, everything else is basically a crap shoot. The #2, #3 and #4 seeds' records are pretty close to identical, but the trend remains that they do better closer to home. #2s, #3s and #4s combined have a record of 28-29 playing close to home, as opposed to 33-47 (.412) when they have to fly.

While I have not done the math of what the 2s, 3s, and 4s would look like if you removed the impact of the games won by 1s close to home, my guess is everyone would come out close to a coin flip.

Pretty much every single year, you can count on a high, favored 1 seed flaming out after flying to their regional. Last year, Miami did the honors in Providence after winning the NCHC title. We all remember BC flying to St. Louis in 2011 only to get obliterated by Colorado College.

What does that mean for this year's bracket?

One #1 seed has the benefit of playing at home in St. Cloud State, who will defend their turf in the St. Paul regional. #1 seed Providence College has a short bus ride to Worcester in their hunt for the Frozen Four.

Quinnipiac also has the benefit of taking a bus to their regional, though they are playing an opponent in RIT who are in their own home state. The final 1 seed, North Dakota, does have to fly to their regional in Cincinnati, putting them on a level playing field travelwise with Northeastern.

Teams playing in state this year across all seeds are BC and Harvard in the Worcester regional, RIT in Albany, and St. Cloud State in St. Paul.

I think these numbers lend credence to the idea that the tournament format should be changed to allow higher seeded teams to host early rounds on campus sites. With travel distance being such a huge factor in tournament performance, teams should have the ability to earn a tournament game close to home, rather than gaining that edge through either paying to host a regional, being a good enough attendance draw that the committee moves them close to home, or just getting lucky.

Since the format doesn't appear to be changing any time soon, keep this factor in mind as you fill out your tournament bracket.