I’ll be using the below statistics and concepts frequently throughout this blog.  Feel free to ask questions in the comments about any of these, but first check here to see if this glossary answers your questions:

Commonly Used Analytics Terms:


Corsi (Sometimes abbreviated as CF% or Corsi On):  Shots on Goal + Missed Shots (MS) + Blocked Shots (BS).  Corsi is the most frequently cited advanced metric.  Corsi is alternatively known as “shot attempts” since it measures not just shots on goal, but ATTEMPTED shots on goal.  Corsi is usually expressed in one of three ways:
Corsi Differential (Raw):  Shots on Goal By your team with you on the ice+ Missed Shots by your team with you on the ice+ Blocked Shots by your team with you on the ice- SOG by your opponents with you on the ice- Missed Shots by your opponents with you on the ice- Blocked shots by your opponents with you on the ice.
Corsi Differential Per 60:  Raw Corsi Differential (see above) while a player is on the ice per 60 minutes.
Corsi Percentage:  The % of corsi events by both teams that are made by a player’s team while that player is on the ice.  So if the two teams have combined for 10 total SOG+MS+BS with you on the ice and your team has 5, you have a corsi percentage of 50%.  If it’s 6 for your team, your corsi percentage is 60% and if it’s 4 your corsi percentage is 40%.

Corsi Differential is the form of corsi used on while and use Corsi percentage.  Extraskater also has corsi differential if you want it.

In short, Corsi is known as a “possession metric” – Corsi % and differential do, over the long term, showcase the percentage of time that the team had possession of the puck – and better teams control the puck in the offensive zone more than their opponents.  Corsi also correlates extremely strongly to scoring chances – so a team that wins the corsi battle over a season will win the scoring chance battle as well.

Fenwick (Sometimes labeled FF%):  Shots on Goal + Missed Shots.  Fenwick is basically the same as corsi, and is presented the same three ways (raw differential, differential per 60, and as a percentage), except it excludes blocked shots.  As such, Fenwick is sometimes referred to as “Unblocked shot attempts.”  Fenwick is at best SLIGHTLY better than corsi since it excludes blocked shots (and players blocking shots does seem to be a skill) but fenwick and corsi are equivalent 99.9% of the time.

Since Fenwick uses a smaller sample size than Corsi, we prefer to use Corsi for individual players or for teams over a short period of time.  That said, for measuring teams over a season, we tend to use Fenwick (in particular, Fenwick during close situations, see below)

Relative Corsi and Relative Fenwick:  Relative Corsi takes the Corsi Differential or Corsi % of a player and subtracts from it from the performance of their team without them on the ice (only in games they were actually playing).  So if a player had a 50% corsi percentage and his team had a 47% corsi percentage when he was on the bench, he would have a relative corsi of 3%.  Relative Fenwick does the same thing.  Relative statistics are used to try and account for how a good (or bad) set of teammates help (or hurt) a player’s corsi or fenwick.   If a player has a bad corsi/fenwick but has good relative numbers, it suggests that the player actually helping the team (he is “driving play forward”), but his normal #s are being affected by how bad his teammates are around him.

Score Effects:  Hockey teams strategically do not play the same all the time.  In particular, when a team has a lead or is behind, they change how they play in a way that can affect statistics.  If a team is down in the 3rd period, it’ll become more aggressive and try to take more shots than they might otherwise (taking lower % shots and shots from positions that might lead to a counter-rush) and when a team is up in the 3rd they go into a defensive shell, allowing opponents to take shots from the perimeter and focusing as much as possible about defending scoring chances instead of scoring.

We call the changes based upon the score “Score Effects.”  It is important to account for these effects if you can (if the sample size is large enough) so Analytics people will often use statistics from either Close or Tied situations.

Close Situations:  The game is considered close statistically when the score is within 1 goal for the first two periods or is tied in the third period.  Fenwick in Close situations is considered the best measure for determining the true non-goalie talent of a team for the most part, as it shows possession skill of teams who do not yet need to change their strategies due to the score.

Tied Situations: Statistics when the game is tied.  Pretty self explanatory.  Occasionally you will see teams referred to by “Corsi Tied” (Corsi % in tied situations) instead of “Fenwick Close”, due to an old article finding it was just barely more reliable than Fen Close.


Quality of Competition (QOC):  Quality of Competition is generally measured in analytics by taking an average of a measure of the quality of opponents while a player is on the ice.  So Corsi Quality of Competition (see gives the average corsi differential or percentage of the opponents faced faced by a player.   Similarly, Time of Ice Quality of Competition gives the average time on ice (or % of his team’s time on ice) of the opponents faced by a player.

Quality of competition over a small sample can matter a lot, but over the course of a season, gaps in quality of competition even between the top and bottom players in the league is very small.  As such, this doesn’t have a huge effect on player #s over a season.

Quality of Teammates (QOT):  Like QOC above, QOT is generally measured in analytics by taking an average of a measure of the quality of teammates a player plays with on the ice. Like QOC, you can measure QOT with corsi, time on ice, or other measures.

Quality of Teammates is a bit tricky to measure, because the performance of a player’s linemates in terms of corsi, fenwick or even +/- is affected by the player’s OWN play.  As such, some sites, such as, only measure the average performance of a player’s teammates when that teammate is not on the ice.  That’s a bit tricky to do sometimes as well, as this can lead to small sample sizes for players who are on the same line all year.  Alternatively, using Time on Ice Quality of Teammates (TOI QOT) avoids this issue.

Quality of Teammates matters over the long run a lot more than competition.  The reason for this is that a team can control what level of players they play with pretty well, while their control over who they face is limited even with the best matching a home team can muster.  So you seriously need to account for this if possible.  Relative Measures (see above) help a little with this, as do WOWY measures (see below).  But it’s tricky.

Zone Start Percentage (ZS%):  The % of non-neutral zone faceoffs that a player is on the ice for that are in the OFFENSIVE ZONE.  While coaches don’t have perfect control over what opponents are facing which of their own players, A coach can control (to an extent) which players take more offensive zone faceoffs and which take more defensive faceoffs.  The greater percentage of offensive zone faceoffs a player gets (the higher his ZS%), the easier it is to score and to put up a higher corsi/fenwick.  The effect isn’t huge, and most players have a ZS% between 40-60%, but there are specialists who are either extremely sheltered (with as many offensive zone faceoffs as possible) or are buried in the defensive zone, who you might want to adjust their #s to account for the ZS%.   A recent study, as well as Neutral Zone Tracking, suggests that each offensive zone faceoff on average results in an increase in raw corsi by 0.4.  Alternatively, people have attempted to adjust corsi/fenwick by only looking at data a short period of time (10 seconds or more) after the latest faceoff.

isn’t a particular stat, but it’s a framework used in multiple sports to adjust for how much a player is improving his team.  WOWY stands for “With or Without You” and it means exactly what it sounds like – It consists of comparing the #s of teammates of a player with that player on the ice and seeing how they compare to how those same teammates play without that player on the ice. has corsi and +/- based WOWYs for players over the last 7 years.  These #s are extremely useful in determining which players are “driving the bus” so to speak and which ones have #s that are mainly the benefit of playing alongside truly good teammates.  It’s obviously not an exact way of measuring, but you can see which guys are truly making everyone around them better.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s