With just over 25% of the Washington Capitals abbreviated regular season now officially in the books, it’s a good time to take a look at several of the the key statistics for the team. Last week we looked at the advanced stats for each of the Capitals defensemen (here), and today we will take a look at the Capitals forwards, more specifically, their ‘expected goals for’ percentage (xGF%) and ‘scoring chances for’ percentages (SCF%).
Expected Goals For Percentage (xGF%)
For those new to the “expected goals” concept, it’s simply a stat that quantifies shot attempts with their associated locations on the ice. The shot locations are then indexed with hundreds of thousands of historical data points that quantify past success rates of shots at that exact location. For example, if a player takes a shot from a location that has a 10% success rate, according to historical data, the player will accumulate .10 for his expected goals score. There are additional attributes involved in the final value, but that’s the general idea in a nutshell.
‘Expected Goals For’ percentages (xGF%) is simply the player’s percentage of ‘expected goals for’ (xGF) in relation to the oppositions ‘expected goals for’ (xGA) when the player is on the ice.
Here are the xGF% values for each of the Capitals forwards through the first 15 games. Six Capitals players have an xGF% greater than 50%, meaning they are creating more expected goals chances than opposing players when they are on the ice. A score over 50% is ideal. [Click to enlarge].
Three top-six forwards (Evgeny Kuznetsov, Jakub Vrána and T.J. Oshie) and three bottom-six forwards (Conor Sheary, Lars Eller, and Richard Panik – the third line) have an xGF% greater than 50%.
Comparing ‘expected goals for’ percentages between players is a basic evaluation metric, typically utilized by many in the hockey analytics community. However, the stat lacks significant context without considering other key characteristics, such as line mates, strength of opposition, score at the time of the shot and many other shot and game-related attributes.
Offensive Zone Starts Percentage (OZS%)
In order enhance to the viability of a basic stat we can add context by supplementing a single stat/data point with other key, related attributes. Key attributes that have a significant impact on a player’s ‘expected goals for’ percentage are those that affect a players ability to generate shots. One of those key stats is offensive zone starts.
One can surmise that players that have more opportunities in the offensive zone will have better chances at generating shots, and thus have higher ‘expected goals for’. Therefore we can compare a player’s ‘expected goals for’ percentage to the percentage of offensive zone starts for the player to get a more realistic picture of a player’s performance.
Offensive zone starts percentage (OZS%) is essentially the percentage of starts for a player that were in the offensive zone, excluding neutral zone and on-the-fly starts. Scoring forwards will typically get a greater percentage of offensive zone starts, for obvious reasons. Theoretically, the more chances they get, the more goals they will score.
This also means they will get more opportunities to generate expected goals and thus, should have a higher xGF% value than defensive forwards who are relied upon to defend, and have more zone starts in the defensive end.
The basic graph below compares ‘expected goals for’ percentages (xGF%) and percent of offensive zone starts (OZS%) for each Capitals forward with more than 50 minutes of total time on ice. [Click to enlarge].
Before we proceed, a few notes on the graphic with regards to specific players. Note that Conor Sheary has the second highest xGF% value among Capitals forwards, but has the 4th lowest offensive zone start percentage. Lars Eller has an xGF% above 50% with one of the lower offensive zone start percentages.
It should also be noted that bottom six role players like Nic Dowd, Garnet Hathaway and Carl Hagelin will have much fewer offensive zone starts, as they are players who are generally placed on the ice in defensive situations.
Relative/Differential ‘Expected Goals For’ Percentage
The chart above does a fair job of graphically illustrating a comparison of a players xGF% and OZS%, but lacks a distinct value to quantitatively differentiate the values between each player. To do this we can simply apply numerical values for the difference between the two values for each player.
The statistics presented and developed above are just one set of data points that detail a players ability to generate scoring opportunities. Another key metric is a players scoring chances in relation to opposing skaters.
A scoring chance (as originally defined by War-on-Ice) is an advanced stat that applies additional attributes to a basic shot statistic, including shot location and shot characteristics (rebound, etc.). Additionally, it is an attempt or shot taken from the “home-plate shaped area” that goes from the top of the faceoff circles, through the faceoff dots, then angled to the goal posts.
Scoring chances are broken down into three categories: low-danger, mid-danger, and high-danger. A player is awarded a “scoring chance for” (SCF) if someone on his team has a chance to score when he is on the ice and awarded a “scoring chance against” (SCA) if the opposing team has a chance to score while he is on the ice. Thus, again, any value over 50% is ideal.
For the purpose of this post we will be including all “danger” scoring chances in a simple comparison with offensive zone starts. Below are the percentages of ‘scoring chances for’ for each of the Capitals forwards with more than 50 minutes of time on ice. [Click to enlarge].
We see similar traits from the aforementioned xGF% graph. Only five Capitals forwards, Conor Sheary, Evgeny Kuznetsov, Garnet Hathaway, Lars Eller and Richard Panik, currently have a SCF% value above 50%, which is not ideal, but impressive for those five players.
Carl Hagelin and Nic Dowd are right at 50%, which is also impressive, considering they do quite a bit of line checking and are often tasked with facing opposing team’s top lines. Extra kudos to Hathaway who is also part of the checking line and has an SCF% above 50%.
Now let’s add the percentage of offensive zone starts for each player as we did for the xGF% stats presented above. [Click to enlarge].
Once again Conor Sheary is shining bright, as he has the 4th fewest offensive zone starts, but the 4th best SCF% among Capitals forwards. Additionally, Lars Eller and Richard Panik have SCF% values above 50%, but have OFZ% that are considerably less than most other forwards. Finally, the checking line of Carl Hagelin, Nic Dowd and Garnet Hathaway, once again shine impressively here.
As we did with xGF% and OZS%, we can generate values to establish the relative and differential values for each of the Capitals forwards.
There are a vast array of statistics that assist analysts in determining how effectively forwards are playing, but a good first measure is determining xGF% and SCF% scores and reconciling their scores with OZS%.
The Capitals bottom six forwards of Conor Sheary, Lars Eller, Richard Panik, Carl Hagelin, Nic Dowd and Garnet Hathaway are doing a really good job of neutralizing opposing teams and generating significant expected goals and scoring chances for the Capitals.
[It should also be noted that bottom six forwards and checking lines deployed in defensive situations will more often than not have positive values in the chart above and listed in the aforementioned numerical tables, as they will always have low offensive zone starts. The numbers are more relative to linemates in this assessment]
They may be somewhat unheralded, and their paychecks may not be as big as top six forwards (Conor Sheary is making $735k and Nic Dowd is making 750k this season), but their value to the team can not be overstated through the first 15 games of the season.
Data courtesy of Natural Stat Trick.
By Jon Sorensen