clock menu more-arrow no yes mobile

Filed under:

A Closer Look at GAR Components for the 2016-17 Red Wings

Ottawa Senators v Detroit Red Wings Photo by Gregory Shamus/Getty Images

In my previous article I took a look at last year’s Red Wings’ performance using Dawson Sprigings Goals Above Replacement Model.

We saw that Jonathan Ericsson performed at the level of a replacement player, while Niklas Kronwall performed below that level. I didn’t go into this, but Luke Glendening was below replacement as well.

Last time I focused on each player’s total GAR as well as how our team stacked up to the rest of the NHL. Today, I’m going to break it down more, since total GAR is made up of six components.

Component Breakdown

The next two charts show how the Red Wings forwards and defensemen performed in each component of GAR. The section to the right of the zero line are components in which the player had a positive number (helped his team). The sections to the left of the zero line represent a negative number (hurt his team). The players are arranged from top to bottom by total GAR.

Here’s what the abbreviations mean:

Draw - Penalties drawn; EVD - Even-Strength Defense; EVO - Even-Strength Offense

FAC - Faceoffs; PPO - Power Play Offense; Take - Penalties Taken

The faceoffs and penalties components are not simply how many faceoffs a player won or lost or how many penalties a player took or drew. If you are interested in reading about how those components are calculated, look at the footnote at the bottom of the article (*)


Data: @DTMABoutHeart: Chart: @pflynnhockey

This and all charts for this article are available in interactive form here. Also, in the charts in this “Component Breakdown” section, the x-axis is not moved out to the NHL maximum and minimum like it is for the other charts. I did that so it’s easy to see the different components. So, even though Zetterberg is at the far right of the chart above, he’s still 10 lower than league-leader Connor McDavid.

I’m going to break down a few of the components below, but a quick look here shows that Henrik Zetterberg, Tomas Tatar, and Gustav Nyquist provided the most even-strength offense. Anthony Mantha, Andreas Athanasiou, and Thomas Vanek also provided solid EVO, although we can also see here the negative side of Thomas Vanek.

Since I’m not going to spend time in this article talking about the penalty and faceoff components, I’m just going to point out a few things here quickly.

Zetterberg and Nielsen provided the most faceoff value to the team. Their 1.6 and 1.8 respectively were above the league average of 0.4.

For the penalties drawn component (Draw), the league average for forwards was -0.1, so any blue bars on the right side of zero are above league average. Justin Abdelkader provided the most value for the team, which is good because without that, his value would be well below replacement level. Dylan Larkin, Darren Helm, and Athanasiou are other positive contributors in that aspect.

On the flip side, the “Take” component measures the impact of penalties the player takes. Probably not surprisingly, Abdelkader is on the negative side of zero here. NHL average for forwards is 0.8. Zetterberg, Tatar, Nyquist, Nielsen, and Riley Sheahan are all above that mark. (Mantha is right at 0.8).

For penalties, since the numbers indicate value for the team, a player with a negative number for penalties drawn indicates that he drew fewer penalties than would be expected from a replacement player. A positive number for penalties taken indicates that he took fewer penalties than would be expected from a replacement player.


Data: @DTMABoutHeart: Chart: @pflynnhockey

Obviously, the Faceoff component does not apply to defensemen. Also, Mike Green was the only Red Wings defenseman with a Total GAR above NHL average. (Klefbom led the NHL blueliners with a 17 total GAR.)

NHL average for the penalties drawn component is 0.0 for defensemen, so Dekeyser, Marchenko, and Kronwall are below the league average. The NHL average for penalties taken is 0.9. Dekeyser and Jensen are above league average, everyone else is below.

For our two lowest defensemen, we can see that different components brought each down to replacement level or below. For Jonathan Ericsson, penalties taken and even-strength offense were the two areas in which he provided less value than a replacement level player, whereas for Niklas Kronwall it was even-strength offense and penalties drawn.

Power Play Offense

Even-strength offense, even-strength defense, and power play offense are derived through a combination of two models: Expected Plus Minus (XPM) and Box Plus Minus (BPM). The links above take you to Sprigings explanation of each.

In our latest podcast, Jay and I interviewed Garret Hohl, who talked about the GAR model. He explained XPM and BPM in a way that I thought was informative, yet easy to understand. The following is what he said about each of these components of the model:

BPM: “ a player impacts the box-score stats, which is your simple stuff, like scoring, hits, shots, turnovers, and takeaways.”

XPM: “basically a super-Corsi that introduces both shot quality aspects and usage adjustment impacts.” **

Power Play Offense is one area that explains why Frans Nielsen was in the same neighborhood as Tatar and Nyquist in Total GAR that we saw last article. We can see on this chart that he was head and shoulders above the rest of the team last year.

Data: @DTMAboutHeart; Chart: @pflynnhockey

Claude Giroux led the league last season with a 4.2. NHL Average was 0.7, so Green was right at the average, while Nyquist, Vanek, and Zetterberg joined Nielsen above league average.

The colors of the bars represent power play time on ice. So we can see that Athanasiou and Darren Helm provided as much PPO as Abdelkader and more than Dylan Larkin with less ice time on the power play.

Even Strength Offense and Defense

I broke this up by forwards and defensemen. We know from last article that the forwards were ranked 9th in the league in Total GAR, so let’s see how these two components contributed to that.


Data: @DTMAboutHeart; Chart: @pflynnhockey

Let’s take the positives first. Recent signing Tomas Tatar was second on the team (and well above NHL average) in both offense and defense. (For this section, the “even-strength” part will be implied so I don’t have to keep typing it.) Nyquist and Zetterberg were both top four on the team in each category.

Frans Nielsen led the team in defense, although his offense didn’t measure up to NHL average. Between Mantha and Athanasiou, Mantha provided more offense without much defense, while Athanasiou was just under NHL average in both components.

On the negative side, while Thomas Vanek helped on the offensive side, his defensive play hurt the team the most of any forward. Keep in mind that this was even with Blashill sheltering him.

As mentioned previously, Abdelkader’s offense was below replacement level. His defense was also below replacement level, although less so. Darren Helm and Luke Glendening are the other two players who were below replacement level in combined offense and defense.


Data: @DTMAboutHeart; Chart: @pflynnhockey

If you remember the last article, you can already guess this isn’t going to be good. Mike Green provided offense closer to the NHL high mark than the NHL average (15th in the league among defensemen). This is good because his defense hurt the team more than a replacement level player.

Xavier Ouellet provided the most value at defense, and his offense was not much below NHL average. Kronwall provided the least value on the defensive side which nearly balances out the value he provided on the offensive side (which still wasn’t that great).

In Closing

In the last article we saw that the Red Wings forwards group did not perform as badly as people might think while the defense did.

I like this model because it can help to see how players can provide value outside of just goals and assists. Someone like Frans Nielsen provides value to a team that can’t always be measured in box score stats. While Thomas Vanek provides solid offense at even strength, other aspects of his game hurt the team.

It’s important to note that GAR does not take into account the value a player brings on the penalty kill. The reason for that is that there’s not a solid understanding of how to measure penalty kill value yet. Finding a way to add this as a component would certainly be helpful in identifying which players help their team in that way.

If you are interested in other player performance models, Evolving Wild recently introduced his model called Weighted Points Above Replacement. I haven’t had a chance to look too closely at the data, but I hope to dive in soon.

* “Both of these “extra” models are derived using the exact methodology originally developed by Sam Ventura and Andrew C. Thomas at War-On-Ice. A player’s faceoff value is determined via a Generalized Binomial Linear Model, that takes into account the skill level of the two players engaged in the face-off. The net value of a faceoff win is set as 0.013 as found by Michael Shuckers. A player’s penalty value is determined via a Poisson Mixed Effects Model, that is position sensitive and gives us the value of a player’s ability to draw and (not to) take penalties. The position adjustment helps account for the fact defensemen tend to take more penalties and draw less relative to forwards. The net value in terms of goals of a single penalty is set at 0.17. This comes from the assumption, each penalty is worth 1.8 minutes of GF60 going from 2.5 to 6.5, GA60 from 2.5 to 0.73. (6.5-2.5 + (2.5-0.73))*1.8/60 = 0.17 “(from Sprigings explanation of the GAR model)

** Even-Strength Offense, even-strength defense and power play offense each use a different weighting of XPM and BPM. If you are interested, here is a chart from Sprigings articles on GAR that shows the weights: