Passing stats: What are they and why should we care? Part one

The public discourse on hockey analytics has grown exponentially in recent years. Much of it has centered on puck possession with terms such as Corsi and Fenwick, shot-based metrics that appear in a variety of forms. At Hockey Prospectus, we would like to introduce another way to measure puck possession: passing statistics.

The idea for tracking passing stats comes from English Soccer. For the English Premier League, sites like Opta Sports and EPL Index have several pass-related stat categories easily available to the interested fan. In the book, Soccernomics, by Simon Kuper and Stefan Szymanski, the subject of data analysis is a recurring theme in the early chapters. The pair interview Gavin Fleig, Manchester City’s head of performance analysis, and he explains that “the top four teams consistently have a higher percentage of pass completion in the final third.”

In hockey terms, that would translate to a team’s completion percentage in the offensive zone. This is where the idea of tracking passes was born. Since both are fluid games, would something as simple as higher completion percentages be a strong predictor of success in the NHL? Even if it wasn’t, what other benefits could there be to tracking passes?

In a series of introductory articles, I will lay out the methodology behind tracking passes, what I have learned along the way, and how they can be used to enhance our analysis of the game. This article focuses on the definitions of what a pass is for tracking purposes, what a completion is, and discusses how pass attempts can be used when discussing possession.

Defining a “pass” and how to record them

In order to begin tracking passes, I had to define what constituted a pass. During a typical hockey game, there are several actions that move the puck from one player to another, not all of which are done intentionally. I settled on trying to capture the intended actions of the player with the puck, as that would eliminate some of the puck luck seen on any given play.

A pass, therefore, is an intentional act of one player to get the puck to his teammate: this can be a pass with the stick, a kick with the skate, or a hand pass where allowed. When one of these occurs, a pass attempt is recorded in the zone the pass originates from. If it is a completed pass, it is also recorded in the zone it originated from, not where it was completed.

Defining a “completion”

Since hockey is a fast game, and decisions and actions are made in a split second, determining whether a pass was completed or not can be difficult. I settled on these three criteria for judging a completion:

1) Maintaining possession;
2) Allowing for the recipient of the pass to make a “hockey move” (e.g. dump in, deflection, another pass attempt);
3) A shot attempt.

Should any of those three actions occur, it is marked as a completion. Again, this goes back to intent on both the passer and the recipient of the pass.

Why record the zone a pass originates from?

This is the most frequent question I have been asked. I decided that as we use shot attempts as a proxy for possession, so too can we use passing attempts and completions for the same purpose. In fact, with hundreds of passes per game compared to double-digit shot attempts, passes give a much more accurate reading of the possession battle within a game. It also identifies where players spend most of their time on the ice (i.e. a heat map of sorts).

Since a player has possession of the puck when he attempts a pass, it serves as a location point for measuring territorial control on the ice.

What is a passing chart?

Completion percentages can reveal things we may or may not already know about certain players. Is a particular defenseman smart with his passing in his own end? Is a forward skilled at controlling play in the offensive zone? Which players link play between defense to offense in the transition game best?

I have been tracking passes for the New Jersey Devils since the start of the 2013 – 2014 season. Here is a look at what a single game passing chart looks like:

(CLICK TO ENLARGE)

Devils_Habs_Game_48_Passing

You will see that completions and attempts are recorded for each zone and a player’s completion percentage in each zone is generated as the raw data is entered. The defensemen and forwards are separated so that one can evaluate the positions both individually and on a group level (i.e. How did the defensemen do tonight?).

At the bottom of each chart, you will see each position’s total pass numbers, as well as the team’s totals for that particular game. In the cells to the right of these totals are percentages that illustrate what percentage of the team’s passes occurred in each zone. This goes back to the tracking by zone idea and how it functions as a guide to possession.

Notable observations

Through 53 games of data, trends began to form. I will delve more deeply into these in future articles, but here is a short preview:

- Using a player’s Even Strength Time On Ice data, we can approximate how often a player attempts a pass;
- Comparing zone pass percentages, we have data to approximate how much time a player spends in each zone;
- Which players are more accurate? Which are the most wasteful with their passing?

A review of passing terminology: this will accompany each article to keep track of the terms that may seem new to many of you.

Pass attempt:  A deliberate attempt to get the puck to a teammate

Pass completion: A pass that results in one of three outcomes:

1) Maintaining possession;
2) Allows for the recipient of the pass to make a “hockey move” (e.g. dump in, deflection, another pass);
3) A shot attempt. When in doubt, common sense will prevail.

Ryan tracks zone exits and passing stats for In Lou We Trust, where he is a contributing writer. He is a long time fan of the New Jersey Devils.

Head_Shot

Follow Ryan on Twitter at @RK_Stimp.

4 thoughts on “Passing stats: What are they and why should we care? Part one

  1. Great idea. I am curious how teams/players currently considered great on possession based on Corsi for percentage 5 on 5 might stack up based on pass attempt/completion figures. Also, how a certain style (say Kings or Blues) might compare with another (say, Hawks or Bruins). I look forward to future posts.

    • Observing certain styles of teams is definitely interesting. The Devils, for example, they typically control the puck down low, pass to the point, and attempt shots. They’re very effective that way, but I think it exaggerates their possession numbers.

      I just tracked the Devils game against the Blues yesterday and territorially, the Devils attempted so few passes in the Blues zone until they were chasing the game down 2-0, yet shot attempts were fairly close on the night; even though the Corsi numbers suggested a close contest, the passing figures clearly illustrate the Blues possession dominance.

      Thanks for reading.

Leave a Reply

Your email address will not be published. Required fields are marked *

*

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>