FanPost

Using Statistics to Predict the NHL Playoffs


Now that the regular season is about over, the hockey-watching universe is turning its attention to the first-round matchups, and the question of who will win the Stanley Cup. At sophisticated places like Fear the Fin, of course, readers understand that simplistic, narrative-driven analyses, like won-lost records, home-ice advantage, or player matchups are often less meaningful than statistics that attempt to measure the underlying quality of a team's hockey. Within the world of advanced hockey stats, no statistic is more frequently cited than teams' Fenwick Close %, or their proportion of Fenwick events (shots on goal and misses) across their various games.

While FenClo% has a strong relationship to playoff success, both empirically and intuitively (i.e., to me, a team's ability to consistently tilt the ice in their favor is a great measure of "good at hockey"), it's obviously not perfect: recent hockey history includes possession-dominant teams like the 2009 Sharks, 2010 Capitals, and 2012 Penguins who failed to advance past the first round, and the 3rd-best FenClo% team this season, the New Jersey Devils, is planning its golf schedule rather than playing postseason hockey. Intuitively, a couple of explanations come to mind. FenClo% is measured at even-strength, so it stands to reason that poor special-teams play could undo good possession play at evens. And, of course, there's goaltending: as was the case with this year's Devils and last year's Penguins, weak play in goal can undermine teams that are "good at hockey", and the 2010 Canadiens can tell you that the converse is true also. So, if I want to understand how teams are playing and make guesses as to what the playoffs hold, I need to set up some metric that incorporates possession, goaltending, and special teams.

So, I made a spreadsheet that listed out each team's FenClo%, even-strength Sv%, PK% and PP%, and because I was too lazy to do some regression-based thing, I decided to experiment with weighted averages of these four statistics. (Note: I downloaded the data for this last Friday, so the season wasn't quite finished.) For example, say you want to see how teams rank if equal weight is placed on possession, goaltending, and special teams (if you're really into the stats thing, think of this as a flat prior); this would mean a weight of 0.3333 on each team's FenClo%, 0.3333 on each team's Sv%, and (0.3333/2) on each of PK% and PP%. This would give San Jose (FenClo% 53.03%, Sv% 92.8%, PK% 84.9%, PP% 20.1%) a weighted average of 66.10; if you're curious, only LA, Chicago, and Boston have higher total scores.

To keep this manageable, I ran five different scenarios: 1) the one with equal weight on all three components, 2) a nod to the Fenwick crowd, with 75% on FenClo%, 12.5% on goaltending, and 12.5% on special teams, 3) a more conservative possession-driven scenario, with 50% weight on FenClo% and 25% each on goaltending and special teams, 4) a special-teams-driven scenario, with 50% total weight on special teams and 25% each on FenClo% and Sv%, and 5) a goaltending-driven scenario, with 50% total weight on goaltending and 25% each on FenClo% and special teams. The results are in the table, below.

Possession 75% Scenario Possession 50% Scenario Neutral Scenario Special Teams 50% Scenario Goaltending 50% Scenario
LA LA CHI CHI CHI
CHI CHI LA LA OTT
NJ BOS BOS SJ BOS
BOS SJ SJ OTT LA
STL STL OTT BOS SJ
NYR OTT STL STL DET
SJ NJ DET DET STL
MTL DET MTL MTL NYR
DET NYR NYR PHI MTL
OTT MTL VAN PIT PIT
NYI VAN PIT VAN VAN
VAN NYI NJ WSH WSH
CAR PIT NYI NYR NJ
PHX PHI PHI ANA ANA
PIT ANA WSH NYI NYI
WPG PHX ANA TOR PHX
PHI WSH PHX NJ PHI
ANA CAR TOR EDM TOR
WSH MIN MIN CGY EDM
FLA WPG CAR MIN MIN
MIN CGY EDM PHX CAR
CGY DAL DAL DAL DAL
DAL FLA CGY TB WPG
COL TOR WPG CAR CBJ
NSH EDM FLA WPG TB
CBJ COL TB CBJ CGY
TOR NSH CBJ FLA NSH
EDM CBJ COL COL COL
TB TB NSH NSH FLA
BUF BUF BUF BUF BUF

Some conclusions to be drawn:

  • It sort of looks like Chicago and LA are the only obstacle to the other reaching the Stanley Cup Final. And Boston: I'll leave it up to you whether you think touching the Prince of Wales Trophy is bad luck.
  • Much talk has been directed at the "paper tiger" qualities of Anaheim and Toronto, but it's interesting to me that Pittsburgh is seldom included in this discussion. For a presumptive Cup favorite, their poor PK and weak possession game has to be a concern. Unless the PDO Fairy continues to sprinkle good fortune on the Penguins, this could be another early exit. Given Ottawa's strong possession game, goaltending, and penalty killing, they should give Pittsburgh all they can handle in the first round.
  • Teams that appear to be clearly overmatched: Minnesota, Toronto, and the Islanders. Teams that may be overmatched: Anaheim (better PP than Detroit, but far inferior possession game) and St. Louis (though I'm really looking forward to this series).

Some caveats:

  • In any small number of games, luck can render any of the above meaningless.
  • The above is based on the full-season statistics for each team, which may not fully represent how they'll be playing now. The Sharks have played well above their season-ending FenClo% since the trades that sent Douglas Murray, Michal Handzus, and Ryane Clowe out of their lineup, and obviously St. Louis' blueline additions at the deadline, and the presumed return of Jaroslav Halak, will help their FenClo% and Sv%.

This item was created by a member of this blog's community and is not necessarily endorsed by Fear The Fin.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Fear The Fin

You must be a member of Fear The Fin to participate.

We have our own Community Guidelines at Fear The Fin. You should read them.

Join Fear The Fin

You must be a member of Fear The Fin to participate.

We have our own Community Guidelines at Fear The Fin. You should read them.

Spinner.vc97ec6e

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9355_tracker