NHL: Behind The Numbers – Possesion Metrics

PROOF IS IN THE NUMBERS: The numbers behind the LA Kings 2011-12 season showed them to be much better than their 8th seed would suggest.


Traditional thinking about the game of hockey contends that it is a spontaneous, free flowing game that doesn’t lend itself well to statistical analysis. After all, there are so many variables at play: on-ice matchups, line-mates, and pure chance or “puck luck”.

There have been quantum leaps in recent years however, in our ability to measure and better understand hockey. Of course, some of these metrics are an imperfect science, but can still provide us with insight into the game that allow us to understand the numbers, what they mean and how to apply them when making selections.

In part one of our Behind The Numbers series, we take a look at Possession Metrics and their implications in betting.


It is generally accepted that if a team is consistently outshooting their opponents at even-strength (and doing so by a wide margin) they are going to win more than they lose. To calculate shot% we take all shots on goal and all goals both for and against, and express them as a percentage. A good team generally will control about 53+% of the shots in a tied-game state.

An expanded version of shot%, which counts up all shots, goals and misses, for and against, and expresses that as a percentage.

In addition to counting up all shots, goals and misses for and against in Fenwick, Corsi also includes all “blocked shots” for and against and is generally expressed as a percentage. Basically any time a shot is attempted, Corsi records that as an “event” for or against. As such, we can read Corsi as another form of time on attack.

WHERE TO FIND THEM: nicetimeonice.com, behindthenet.ca

It is important to judge teams based off their performance controlling events in a “tied-game state” as teams regularly “play to the score”. Meaning a team that is up by two or three goals will be generally not push the play and will be outshot. Compounding this is that these teams will get outshot at an increasing rate as the game nears completion and the score gap remains.

You generally need 30 games of data to reliably use Shot% Tied, Fenwick% Tied or Corsi% Tied to tell what a team is really made of. If you looked at the Fenwick% Tied of the LA Kings between the trade deadline and the end of the regular season (basically once they acquired Jeff Carter), they were the best club in the NHL, and proved it by steam-rolling through the playoffs, going 16-4, en route to their first Stanley Cup.

Of course, Fenwick Tied% isn’t an all-encompassing, perfectly predictive metric. The Pittsburgh Penguins were second by this measure last season and were out in the first round. The Bruins were middle of the pack in 2010-11 and went on to win the Stanley Cup because no one could score on Tim Thomas.

It is important to remember that Fenwick, Corsi and Shot% Tied doesn’t factor in things like “special teams play”, or more importantly, like the Boston Bruins example above, the quality of goaltending a club receives. As anyone who knows hockey can tell you: nothing can ruin your day as quickly as a hot goaltender. But with all things being equal, the teams that are the best at controlling the events in a game, particularly shots (i.e. scoring chances) when it matters (in tied or one-goal games), are the teams that win the most.

One comment on “NHL: Behind The Numbers – Possesion Metrics

  1. Pingback: NHL: Behind The Numbers – Putting It All Together | PlayNow Sports Blog