Now that the Lightning have played a little over three full weeks of the schedule, we’re introducing a new weekly feature here at Raw Charge called the Weekly Report. Businesses use data to analyze their performance and assist in future decision making. They identify Key Performance Indicators (KPIs) and check on those KPIs at regular intervals.
Since it works well in business, I’m going to take this approach and apply it to hockey. After each week, I’ll compile a report that highlights how the Lightning are performing as a team and per individual so that we can review the gameplay of the past week and make predictions about what should happen next. This is our first installment.
The first page of the report shows several team statistics at 5v5. The blue charts on the left side of this page are what we’ll call “process” metrics. The gray charts on the right side are what we’ll call “results” metrics. Process metrics help us understand how well the Lightning are playing. Results metrics tell us what the outcomes were.
As an example, shot generation and shot suppression are the best publicly available indicators of how well a team is playing. If a team generates more shots, we expect them to score more goals. If a team gives up fewer shots, we expect them to give up fewer goals. But because hockey is a dynamic game with so many moving parts and a healthy dose of randomness, good process doesn’t always result in good outcomes.
This is the type of information we’re looking for from this view. Did the Bolts have good process but bad results? Did they have bad process but good results? If either of those things happen, we can adjust our expectations for the future. Good process should eventually lead to good results. And bad process should eventually lead to bad results. So if we see something other than that happening, we can be prepared for what might happen next. All data in this article comes from corsica.hockey and is score and venue adjusted.
The first thing I noticed in these charts is the overwhelming averageness of the Lightning’s play so far this season. They rank 15th in the NHL in both shot generation and shot suppression, putting them exactly in the middle of the league by both measures. While we are still in the early part of the season, the numbers here do indicate that the Lightning are not performing at the level they would expect if they hope to make another deep playoff run. That said, they have plenty of time to reach that level.
The only area where we see a significant gap between the process and results as discussed above is in the goals against number. Despite being basically league average in both shots against and expected goals against, which is a statistic that uses the location, type, and other characteristics of shots to estimate their likelihood of becoming a goal, the Lightning are 28th in goals against. The reason is fairly easy to identify as the goaltending has not been great thus far (ranking 26th in save percentage). With two skilled goaltenders, the Lightning’s team save percentage should increase as the season progresses and as that happens, the goals against should align better with the underlying defensive statistics.
The next page of our report looks at special teams play. Logically, we only need to look at offensive metrics on the power play and defensive metrics on the penalty kill because play generally goes in one direction during a power play. Special teams performance is much more difficult to quantify than 5v5 performance because teams spend so much less time at 5v4 or 4v5 than at 5v5. And because of that, the sample of events is smaller, which allows for more randomness.
While the Lightning are average at 5v5, they have been excellent on special teams. They rank in the top ten in every metric presented here including goals for per 60 minutes on the power play, and goals against per 60 minutes on the penalty kill. If the Lightning can maintain this kind of performance on special teams and reach the level of 5v5 play they’ve shown over the last two seasons, they will once again be one of the most formidable teams in the league.
Now that we’ve looked at the team performance, let’s look at each individual player. For the team as a whole, we used 12 different metrics to assess performance. If we use 12 different metrics to assess 18 different players, we would need to look at 216 different graphs. I have an extreme fondness for graphs. But not even I want to try to scroll through 216 graphs and process what they are telling me. So instead, we’re going to start a single metric to get a high level view of how each player is performing. The metric is called Game Score. Dom Luszczyszyn created it and his explanation of it can be found here. But as a summary, game score is a box score statistic that takes into account a variety of individual and on-ice statistics such as shots, assists, and goals both for and against and gives each player a score for each game. Ryan Callahan and Slater Koekkoek are not included because neither has played enough games to have a meaningful sample.
Using Dom’s game score, we can see that four players are the most responsible for driving the Lightning’s success at 5v5: Steven Stamkos, Nikita Kucherov, Anton Stralman, and Victor Hedman. These four are the only Lightning players who have consistently performed at a high level this season. Alex Killorn and Brayden Point have also played fairly well. But outside of that, the Lightning players have largely underperformed compared to what would be expected at 5v5. The defensive pairing of Jason Garrison and Andrej Sustr is particularly concerning with both defenders currently performing below expected in the bottom fifth of defenders in the NHL at 5v5.
For additional context, the final page of our report uses relative expected goals-for percentage to help us understand how each player has impacted the team’s performance. Being near the top of this graph suggests that the team plays better with that player on the ice while being far to the right indicates that player is scoring at a high pace. So while Stralman hasn’t scored a primary point at 5v5 yet this season, the team is much more likely to outscore the opponent while he is playing.
Based on the charts above, one could easily conclude that Jonathan Drouin has not performed particularly well this season. But as a reminder, these numbers are only 5v5. Drouin has been highly effective on the power play creating a dynamic combination with Nikita Kucherov. Because players play so few minutes on the power play, it can be difficult to meaningfully measure that performance with statistics this early in the season.
As the season progresses, I will work on incorporating more information about how individual players are impacting special teams. And given how much the Lightning have relied on their special teams so far this season, that will be an important piece to add to this analysis.
As a final note, keep in mind that because the season is so young, the numbers above still have limited meaning. The next game will increase our sample by 8.3%, meaning that we can still expect to see big changes in the numbers from game to game. But as we get further into the season, the statistics will stabilize and we’ll be able to identify nuanced changes that will help us learn more about how the Lightning are playing.
I encourage you to offer your own interpretations of the reporting in the comments. Making a report is fun for me but it only becomes valuable if we can use it to learn more about our team.
