The Tampa Bay Lightning started the season by killing 28 consecutive penalties over parts of seven games. That streak ended on Sunday night when Chicago scored on the power play late in the third period.
Special teams are susceptible to misleading results because of the small sample sizes where streaky shooting or goaltending can make a team look terrible or great regardless of how they are playing. The Lightning’s streak raises the question of whether this was a stretch of excellent penalty killing or a run of good luck.
If you ask people who follow the Lightning closely to list the team’s primary weaknesses last season, many would point to the penalty kill. In the offseason, the team did not retain the coach in charge of the penalty kill and defense, Rick Bowness. They instead hired new assistant coaches and moved Todd Richards from focusing primarily on forwards and the power play to the defense and penalty kill.
But the data didn’t always support the narrative that the penalty kill was a problem. All numbers that follow in this post are via Corsica.
The Lightning finished the season with with the 6th best expected goals against at 4v5 in the NHL. That doesn’t sounds like a team where the penalty kill was a problem. Despite that, they finished 24th in the NHL in goals against on the penalty kill. A big gap between expected goals and actual goals immediately points to goaltending. And in terms of the difference between expected and actual save percentage, the Lightning had the worst 4v5 goaltending of any team last season.
But even that seems like a simplification. While the goaltending was part of the problem last year, I don’t think blaming that and calling it a day captures the whole story. To look deeper, we’re going to view the Lightning’s penalty kill over the last 89 regular season games including all of last year and this year. I’ve chosen to focus on this range because the composition of the roster, particularly in net, was significantly different prior to last season.
The first group of charts focuses on how the skaters performed and ignores goaltending entirely. The first two charts are the separate components of the third chart. Shot quantity is how many shots the team allowed compared to the rest of the league. Shot quality is how dangerous those shots were. And expected goals is the output we get when we think of how many goals the Lightning would expect to allow based on the number and quality of shots they gave up at 4v5.
The measure on the y-axis is standard deviations from the mean for each metric and the lines are smoothed. I’ve done this to make the trends more readable because of the small sample sizes of 4v5 time. In all three metrics, zero is average, high is bad, and low is good. The blue line is the Lightning and the gray lines are the other 30 teams.
The first thing worth noting is this chart is that the Lightning have consistently allowed an average amount shots at 4v5 over the last season plus. That is not true of shot quality. The danger of the shots they permit against their goaltenders has fluctuated wildly. They have at times been the best team in the league and at times been the worst. That’s quite a swing.
And as expected, the expected goal chart mimics the shot quality chart almost exactly. If the team consistently allows an average number of shots, but a wild variance in shot quality, then the expected goals will be driven by the shot quality.
The takeaway from this is that while the team ultimately finished the year with strong expected goal numbers, they did not play well all year. And at times, they were downright terrible. That gives a little insight into why a team with such good overall numbers might be perceived to struggle.
It also isolates the cause of the team’s problems. When the Lightning perform poorly on the penalty kill, allowing more shots isn’t the cause. Instead, they go through stretches where they bleed dangerous chances and put excessive pressure on the goaltenders to make saves.
To this point, we’ve ignored the goaltending. But the next step is to include that. Just like the previous set of three charts, the first two metrics below combine to give us the third. We start where we finished in the previous set with expected goals. If we factor goaltending in with the the expected goal numbers, we arrive at actual goals.
Here we see that yes, the goaltending was below average for most of last season. The particularly bad stretch of goals allowed on the power play coincides with the worst of the skater play in terms of expected goals and the worst of the goaltending. That makes obvious sense. If the skaters are bad and the goalie is bad, the penalty kill is going to allow lots of goals.
So what do we make of the recent hot streak on the penalty kill? The Lightning’s start to the season is a mirror image of the worst point of last season. They’re still allowing the same amount of shots. But like we discussed above, they’ve done an excellent job of suppressing shot quality. And that leads to suppressing expected goals.
Coinciding with the stretch of good play by the skaters, the goaltenders have hit a hot streak in net. Just like bad skaters and bad goalies make for bad results, good skaters and good goalies make for good results. And the Lightning penalty kill is suppressing goals at as good a rate as any team over the last two seasons.
To get back to the question we posed at the beginning of the article, this is more than just a lucky run. The Bolts have performed well at 4v5. The goaltending won’t be this good forever but even average goaltending with a penalty kill minimizing dangerous chances against like this would achieve good results.
Suggesting that a seven game stretch can tell us anything about the rest of the season would be absurd. Especially on special teams. What we know is that the Lightning skaters have played this well on the penalty kill before. They goaltenders have also played this well before. They just haven’t both played this well at the same time. So this isn’t a total fluke.
The goaltending will inevitably regress. The question is whether the coaches can figure out what drives the fluctuations in shot quality. Looking at big picture numbers like this can only tell us WHAT is happening. But it can’t tell us WHY. Answering that would involve tape study and digging into more granular stats. During the bad times, maybe they allow too many clean passes. Maybe they fail to clear the net front. Maybe they cede too much space and don’t apply enough pressure.
The causes could be myriad. But the key for the coaches is isolating the things that make them successful in suppressing shot quality and getting the players to execute in those areas consistently. Because if they can do that, they can eliminate or at least mitigate the poor stretches that plagued them last season. Provided of course that the goaltenders can do their part.