• RSS

In hockey, teams are often given labels to identify their style of play.

"Team X is a hard-working, grind-it-out team."

"Team A is a puck possession team."

One of those characterizations can be quantified with data, and it's the latter.

Advanced statistics in hockey are becoming widely popular and have become more refined over the past few years. Instead of tracking goals (as plus-minus does), possession metrics take note of shot attempts for and against, and in doing so, can show patterns and trends when certain players are on the ice.

In tracking teams do when they have the puck, we can determine things like puck possession, scoring chances, how teams spend their zone time, and more.

Puck possession has always been important in hockey, but analysts have proven how valuable it can be. Shot attempts are the best proxy for understanding and quantifying puck possession.

The idea for quantifying puck possession through shot attempts originally came from Tim Barnes, who posted his work online under the pseudonym Vic Ferrari. Barnes, who now works for the Washington Capitals, coined the term "Corsi," naming the stat after Buffalo Sabres goaltending coach Jim Corsi.

A shot attempt is counted any time a player tries to shoot the puck. They are counted as a shot on goal, blocked shot or missed shot. By adding those three types of a shot together, you have the number of shot attempts.

Comparing the number of shot attempts for and against a team helps determine which teams are best at controlling the puck. This metric can be applied to individual players, based on the team's shot attempts for and against while the player is on the ice. These statistics are typically used in even-strength situations because the majority of a game is played at even strength, and it affords the best indicator of how players or teams perform on a regular basis.

This can be expressed as raw shot attempts for and against totals, as a whole number differential where the calculation is shots attempts for minus shot attempts against, or a percentage where the calculation is shot attempts for divided by total shot attempts for both teams.

There's a strong correlation between shot attempts percentage and winning hockey games. Which team has the most shot attempts might not be a deciding factor on any given night, but teams that control puck possession the best have a higher probability of future success.

This is very similar to shot attempts, but with one caveat: Blocked shots are not counted. Shot attempts are three types of shots (shots on goal, missed shots and blocked shots), and unblocked shot attempts are two types of shots (shots on goal and missed shots).

The concept, which was first developed by former Battle of Alberta blogger Matt Fenwick, is based on considering blocked shots to be a skill, and not as random from team-to-team as other events on the ice.

All of the same ways that shot attempts can be quantified (as a raw total, as a whole number differential, or as a percentage) remain the same and the calculations are the same.

Shot attempts percentage and unblocked shot attempts percentage are usually pretty similar, but differences between the two can illustrate which teams are better at blocking shots at their end of the ice and avoiding having their shots blocked at the opponents' end.

"Puck luck" is a term that's used with ambiguity in hockey, but by adding on-ice shooting percentage and on-ice save percentage (SPSV%, also known as PDO) it gives us a statistic that measures that concept. The idea was developed by Brian King, who began using it in comments on Barnes' blog under the username PDO.

On-ice shooting percentage and on-ice save percentage are a team's shooting percentage and save percentage. At the team level, that is simple. At the player level, it is the team's shooting and save percentage when a certain player is on the ice.

An average SPSV% for a team or player is expected to be 1000. If a team saves 92.5 percent of the shots it faces and scores on 8.2 percent of the shots it takes, that equals a SPSV% of 100.7, which will be expressed on NHL.com as 1007.

If the number is above 1000, it means the team or player is going through somewhat of a lucky stretch and is likely to regress. Likewise, any team or player below 1000 is likely to see an uptick to push that number toward the mean.

In large sample sizes, SPSV% can help weed out teams that may be over or underperforming. The 2013-14 Colorado Avalanche were a good example. Colorado finished first in the Central Division despite finishing 25th in shot attempts percentage (46.9). One of the major factors in was the team's SPSV%, which was third-highest in the League at 1018.

Over a longer stretch of hockey, the Avalanche would have expected to see that number slide closer to 1000, meaning more goals allowed, or fewer goals scored. This season, the Avalanche are still preventing goals well, but the offense has dried up and they are outside the Stanley Cup Playoff picture.

Some teams, like the Boston Bruins, have been able to sustain a higher SPSV% over time with strong goaltending across multiple seasons, but for the most part teams hover around the 1000 mark.

A coach is unable to choose which players he sends out for a faceoff after his team ices the puck, but there are plenty of other situations when he makes those decisions. And who a coach chooses to deploy in certain situations can offer insight about strategy or willingness to trust a player's defensive acumen.

Bruins center Patrice Bergeron is a great example for this concept. A two-time Selke Trophy winner as a forward who best demonstrates skill in the defensive component of the game, Bergeron is a shutdown player.

Last season, he was put on the ice in the defensive zone for 33.8 percent of his shifts that started with a faceoff. This was tied for the 23rd-highest number among forwards. What's most impressive about Bergeron is pairing this information with his shot attempts percentage, which was highest among NHL forwards at 61.2.

Players who start more shifts in the defensive zone are not expected to help their team dominate the puck possession battle, but someone like Bergeron who can is incredibly valuable. There are forwards who fall on the other end of the spectrum, either because of a lack of defensive accountability or because of their individual or team's offensive prowess.

A majority of these statistics are viewed primarily in even-strength situations; there are other modifying terms and filters used to gain a further understanding of certain areas of the game. One of these concepts is a "relative" statistic; in this example, shot attempts relative (formerly known as CorsiRel). What this statistic shows is how a player is performing relative to his team's average.

Other modifiers for statistics relate to score situations. Hockey games aren't played the same way across 60 minutes. When the score begins to change, teams intentionally or unintentionally adopt different strategies.

When ahead, a team may play a more conservative game, leading to fewer shot attempts, opposed to when a team falls behind it may play more aggressively. The four different score situations are even (tie game), ahead (team is leading), behind (team is trailing), or close (within a goal during the first two periods, or tied in the third).

Going off the idea that when a game is tied or close is likely when teams will play at a level consistent with their true ability, many people feel observing games in those score situations offers the best indicator of how well a team is actually playing. Last season, the Kings led the League in shot attempts for percentage in score-close situations at 57.3 percent.