Two: Where to Find NHL Stats, A Novice’s Guide

Behind the Net

The grandaddy of the advanced stats websites, behindthenet (curated by the great @Hawerchuk of Arctic Ice Hockey) lacks the user-friendly atmosphere of extraskater, but, for the most part, it offers the same basic information. Because it can be a rather cumbersome user experience, let’s go through a few typical ways to get information from the site.

While there’s a variety of ways to use the site, the two most common functions I use are the “Player Breakdown” and “Find Player” options under the “Statistics” button.

With “Player Breakdown” you can sort various groups (the whole league, an individual team, just defenders, etc.) according to various criteria. Most of what you will look for at behind the net is available here.

The site is highly interactive (meaning you can isolate for a large number of variables) and can seem overwhelming. Here’s the “Player Breakdown” page unfiltered.

Already, we can see that this is a pretty hectic page of information. It doesn’t all fit on the page and there’s a lot more we can add via the filters (use the “toggle views” and “more filters” buttons to further expand or narrow your search). Here’s some tips on how to use it effectively.

Usually, when you are going to BTN you are going for something specific (be it a player’s #s for a given year, a team’s ranking by a certain stat, etc.). So, the first thing I do is bracket off extraneous information. Two quick ways to do so are by isolating the team you want and setting a games played threshold. This gets us here:

In this red box is the vast majority of the advanced stats information you will want for a given team (TOI/60, CorsiOn, CorsiRel, Corsi-based QualComp and PDO). Unless you are looking for something specific keep your focus here. A nice feature of BTN is that whatever stat you order the team by (in this case CorsiOn), the site gives you a short, easy to understand definition.

BTN also offers a second quality of competition (and teammate) option, one based on the +/- of the opposition. You can pull it up by selecting “QualComp” in the “Toggle Views” section.

At both the team and the league level, I often use BTN for point per 60 stats. These are scoring rates in various on-ice situations that equalize the players regardless of playing time/opportunities. In order to filter out noise, I uncheck everything but “scoring” in the “Toggle Views” section.

Make sure you set a games played threshold in order to filter players with limited games in hand. This can also be used to compare players across the league or to look at power play scoring rates equalized for playing time.

The “Find Player” section is used to buckle down a single player and look at their #fancystats over their career. It gives you a chance to see if a player, for example, is improving in their Corsi numbers. Counterintuitively, the players are listed alphabetically by first name. Here’s Taylor Hall’s page:

Some Kind of Ninja

Somekindofninja, created by Greg Sinclair, primarily offers two services, both of them visual.

The most commonly used service is simply to graph all of BTN’s data into player usage charts. Those familiar with extraskater’s usage charts will find these operate on similar guidelines (note: extraskater used a TOI-based QualComp measure that differs from BTN’s). In general, usage charts give a thumbnail sketch of a huge amount of information (Vertical = QualComp; Horizontal = Zone Starts; Color = Corsi; Size = (in this case) how good/bad your Corsi is, often size can refer to TOI). Here’s a chart for Edmonton from last season:

Alongside standard features like setting a team, a games played threshold and on-ice situations, these charts allow you to toggle from CorsiOn to CorsiRel and from QualComp to QualCompRel and from QualTeam to QualTeamRel.

But, the crucial feature this site makes available in my opinion is graphing a player’s entire career on one chart. This allows you to see how a player has been used and performed across multiple seasons. Here’s Taylor Hall’s graph:

Super Shot Search

Some kind of ninja’s other service is to provide a graph of the shot locations (shots and goals) of all registered shots in NHL games. This service gives you a sense of the “shot distance” and location of all shots for or against a given team. This information can be filtered in a variety of ways. For example, here’s a chart of all the shots taken by the Oilers while Taylor Hall was on the ice at even strength last season:

Rob Vollman’s Player Usage Charts

The pioneer of the usage chart, Rob Vollman, maintains his own website with a very helpful tool for charting a large variety of scenarios (his site hockeyabstract also provides links to all his analytics writing, itself an invaluable resource).

In order to orient yourself with the usage charts you can watch this instructional video and check out this handy visual provided by Vollman and Co.

(Source: Aaron Nichol in Vollman’s 2011-12 Player Usage Charts)

Basically, this rough schematic allows you to quickly translate a player’s position on the chart into a qualitative judgment about how a coach is deploying a player.

Here’s a chart of the Oilers’ D that played at least 40 games last year:

By and large, using the chart is fairly intuitive (despite the army of options available to you, it’s not hard to navigate) and the video link above explains them in great detail. I would simply like to point out to the novice a few things to get them going without too much effort.

For some reason, Vollman has the chart set without displaying the player’s names. Be sure to turn the naming feature on. Also, be sure to set a threshold for games played to bracket off irregular players. The rest of the chart works like every other usage chart you’ve encountered more or less, except that Vollman has handily added another feature: you can include points per 60 stats in the chart (these will only be visible in the table beneath the chart or if you hover your mouse pointer over a player’s bubble).