clock menu more-arrow no yes

Filed under:

So who's going to win Lord Stanley's Cup?

New, comments

It's possible to do a lot of things through statistical analysis.  One of those things is predicting the most likely outcome of any given single game or any given playoff series.  With hockey, this is a simple task because goals are relatively rare (unlike, say, basketball) and they all count for the same number of points (unlike basketball or football).

Brian Burke from Advanced NFL Stats already did the hard work for me in this case by determining that hockey mostly fits what's called a "Poisson distribution".  Once you're done having your brain fried by the impenetrable Wikipedia article, come back.

Okay, provided your skull is still intact, I'll continue.  Poisson distributions are useful when you're talking about relatively rare things, like goals, that happen without regard as to what came before it.  What I mean is that if the Penguins score a goal, that doesn't make them any more likely to score the next goal than it already was.  If you flip a coin and it lands on heads, it doesn't affect the result of the next coin flip.

So, given that we know how many goals every team scores and allows per game on average, we can determine the most probable score of any game, and we can use that information to determine the likelihood of a team winning a seven-game series.  There's a lot of math I'm leaving out, but that's okay; I'm not writing a textbook, I'm making a post on a hockey blog.

On with it, after the jump.

Eastern Conference, first round.  We've got Washington matching up with Montreal, and Washington has a 85.1% chance of advancing.  In the New Jersey series against Philadelphia, New Jersey has a 58.8% chance of taking the series.  I think if you're a Penguins fan, you should hope things don't shake out that way.  For Buffalo and Boston, Buffalo has a 61.2% chance to win a seven-game series.  Pittsburgh and Ottawa?  You can rest easy, as our Penguins have a 62.7% likelihood of winning.

Second round, that leaves the top four teams, which seems fairly intuitive given the teams involved.  Washington against Pittsburgh doesn't look too inviting, what with Washington's 77.2% chance of victory.  A New Jersey-Buffalo series will be much closer, basically a coin flip.  The Devils have the slight edge, though, at 51%.

Washington should cruise past New Jersey to take the Eastern Conference with their 73.1% chance of prevailing.

So what about the West?  In the first round, San Jose has a 67.2% chance to take out Colorado, Chicago has a 76% chance to beat Nashville, Vancouver has a 62.8% chance of beating the Kings, and Detroit has a 52.4% chance to upset Phoenix.  Interesting how the statistical probability reflects what most people consider to be the likely outcome of that series.

In the second round, we'd have a San Jose-Detroit matchup that the Sharks have a 65.8% chance of winning, and Chicago has a 53.4% chance to take out Vancouver.

Chicago, facing San Jose, then has a 51.3% chance to end up as Western Conference champions.

Washington-Chicago would be an epic matchup given that it would be a rematch between Alex Ovechkin and the team he hindered with the hit on Brian Campbell.  It would also be the third straight trip to the Stanley Cup Finals for Marian Hossa and his third straight loss, as Washington has a 62.8% chance to best Chicago in a seven-game series.

The folks over at Japers' Rink shouldn't start planning a parade route just yet, though.