Anders Lee’s Playoff Performance: On Big Stakes, and Small Sample Sizes

Truth be told, it was hard to come up with a focus topic this week. The Islanders are down 2-0 in the second round to the Carolina Hurricanes, and the common forward-thinking theme is that the Islanders need to win Game 3. Of course, that is true, as the number of teams that have come back from a three-game deficit are less than the fingers I have on one of my hands.

The Islanders have played really well through the first two games of the series. Both home games were decided by one goal, and there were a total of four goals scored over the course of 6+ periods. For the Islanders, that they are down by two games is simply a matter of flipping tails twice when the bet was on heads.

To quantify that, the Islanders are controlling the share of expected goals, share of scoring chances, share of high danger chances, and still have a total of one goal scored through all of that. By the way, that goal was scored on the power play, which means the Islanders’ 5-on-5 shooting percentage in this series is 0.0%.

Shooting ruts happen, and they are unpredictable. There’s no telling when they will start, when they will end, or how high stakes will be when it happens. It can be fickle and it can seem unfair. But in a game like hockey, which contains a very high amount of variance across the board, it’s a reality that all players and teams live with.

That brings us to Anders Lee, who has hit a goal scoring lull at a pretty inopportune time. Soon to be an unrestricted free agent, Lee showed he could be plenty productive without John Tavares, finishing with 28 goals and 51 points during the regular season. As referenced in my “Extension Talk” piece, Lee also finished sixth in the league in 5v5 individual high danger chances during he regular season. He’s currently tied for 8th in that metric during the playoffs, too. But he only has one (empty net) goal in six games, which has led to criticism and even questions of if he is worth re-signing.

So, we’re going to dive in.

The story of Anders Lee is not just an even strength one. His ability to create space by using his body in front of the net on the powerplay is admirable. In part, that’s led the team to a better expected goals output than one would expect when actually watching the team struggle to score in higher leverage situations. So for the rest of this piece, we’ll be looking at all situations metrics to really hone in on what’s going on with Lee.

This first chart is a five-game rolling average of Lee’s individual shot attempts per hour over the course of the season. In other words, what this shows is for every five games, how many shot attempts is Lee taking himself on net in any situation (even strength, power play, empty net etc.). Like we would expect, there are some ebbs and flows here, but since the playoffs have begun, his attempt rate is near the highest it has been all season. In fact, his 18.12 attempts per hour after Game 1 of the Carolina series was his tenth highest mark of the season.

Lee ranks 41st in individual shot attempts per hour among all playoff players, That’s translated well in his overall on-ice metrics, as the Islanders have taken 53.92% of attempts when Lee (alongside Mathew Barzal and Jordan Eberle) is on the ice, score-adjusted at 5v5 play. That is +9.59% relative to his teammates, making that line one of the best in the league during these playoffs in terms of getting pucks to the net.

Of course, that is all great, but how about Lee’s bread and butter: high danger chances?

In the above chart, we’re still using five-game rolling averages but the area chart is now blended with all attempts (blue) and high danger attempts (orange). In other words, the orange area of the chart equates to Lee’s high danger chances and the remaining blue area is all of his other attempts.

To note, high danger chances are largely defined by attempts taken in the area in blue below.

As expected, there are similar peaks and valleys throughout the season, but similar to the overall attempts chart, individual high danger chance output started to ramp up just as the playoffs began. Lee’s current five game rolling average (after Game 2 versus Carolina) are his ninth highest total of the year (11.00).

It gets even better when you look at Lee’s metrics relative to the league where he ranks first in individual high danger chances for per hour during the playoffs at 11.54. That’s also two more high danger chances per hour than the second place player (Brian Boyle; 9.48), which depicts an even brighter picture in terms of Lee’s ability to get to the net and create scoring chances for himself and the Islanders. To add to this, we can take a look at expected goal share. When Lee is on the ice, the Islanders have a 62.29% share of expected goals at 5v5 score-adjusted play, which is quite good.

But why is he not scoring?

Good question, and ultimately, so much of that comes down to a variety of factors.

The first one is a bit of luck. Using Evolving-Hockey’s expected goal model, above are four of the top 20 attempts most likely to become goals from Game 2. All of these came off of Anders Lee’s stick... and really in the high danger position of where is “office” lies.

Moreover, shooting percentage functions with extreme volatility, especially in smaller samples (such as five-game rolling averages). The reason for that is simply because of the nature of the metric. Players take a handful of shots every five game, and the conversion rate of those shots varies due to where the shots are taken, the effectiveness of the defense and goalie, among a plethora of other variables.

Looking above, the chart blends shooting percentage on top of the previous area charts that show individual shot attempts and high danger chances. The black line in the chart represents shooting percentage, and as we can see, the fluctuations are a lot more intense - even hitting zero for games at a time.

What we see is somewhat of a juxtaposition where Lee is clearly creating more chances than normal for himself, but is shooting at a lower rate than his average is. Obviously, the timing on that is awful considering the high importance of the playoffs, but it’s clear from the above chart that it’s also not abnormal for shooters to slump.

Put even simpler, Lee’s regular season all situations shooting percentage was 13.7%. During the playoffs, it is just 5.9% and the one goal, as noted above, was scored on an empty net.

So what’s this all mean?

At the end of the day, the timing is bad given the stakes, and the luck factor has been indicative of the first two games for New York’s fortunes. Small sample sizes such as the Stanley Cup Playoffs should not dictate, in isolation, the decision to keep or let go of a player. And in the case of Lee — it really should not have a bearing considering his importance to the team as captain, the desirability of the Islanders for free agents, and his productivity. Otherwise, you may end up making a mistake like this:

But certainly, there’s no getting around the fact that it is problematic for the Islanders if their leading goal scorer is not putting the puck in the net. Still, if you believe in process over results - as Barry Trotz has repeated ad nauseam - there’s not much to do but hope things take a turn for Lee, as his production is vitally important to the team.

And if you’re going to place a bet on a player to turn it around, it probably makes sense to place a bet on the league leader in high danger chances and the number two ranked player in expected goals for.

The stakes call for it.

Data from this piece from NaturalStatTrick.com and Evolving-Hockey.com