San Jose and the unsustainable penalty kill

The San Jose Sharks’ penalty kill is a bit like Eminem’s “Lose Yourself.” On the surface it looks calm and ready (posting an 85.2% kill rate, good for second in the league as of this writing), but if they don’t address the process behind that kill rate, the clock on their good fortune will run out — time’s up, over — blaow! — and they’ll snap back to reality, regressing back toward what their PK more accurately resembles.

According to Naturalstattrick.com, San Jose allows the 15th-most shots against (per 60 minutes) on the penalty kill. They allow the seventh-most shots on goal against. Finally, San Jose allows the eighth-most high-danger chances against. Corsica.hockey’s expected goals against model isn’t as damming: It shows the Sharks as allowing just the 14th-fewest expected goals against per 60 minutes of penalty kill time. Still, those numbers are average at best and certainly not shot-prevention numbers that lead to sustained penalty kill success.

You can see this visually, thanks to Micah Blake McCurdy at hockeyviz.com. The following chart diagrams the unblocked shots San Jose allows on the penalty kill. Dark pink/magenta represents more shots allowed compared to league average and green represents fewer shots allowed.

While San Jose’s penalty killers do a fairly good job limiting shots from the outside, they are allowing plenty of dangerous shots right in front of the goal.

The real reason San Jose’s penalty kill appears better than it really has been is their goalies. Since the beginning of the 2014-15 season, 44 goalies have played 82 goalie seasons with at least 200 minutes of 4v5 ice time. The chart below shows the mean and +/- 2 standard deviations from that mean of that sample size of 82 goalie seasons. It looks at: Save percentage (Sv%), the difference between actual and expected save percentage (dSv%), Low-, medium-, and high-danger save percentage and goals saved above average (GSAA). Below the historical figures are Dell and Jones’ numbers so far this season.

Anything highlighted in red is more than 2 deviations above the historical mean. Anything in orange is between 1 and 2 standard deviations above the mean.

You can see that all of Dell’s numbers are above what you could reasonably expect given the historical information we have. Jones is outperforming a few of the historical categories, too. Given the shots they’ve faced, Dell is expected to have a save percentage closer to 88 percent; Jones, closer to 85 percent. If both goalies had been allowing goals at their expected rates, San Jose, at 4v5, would have allowed about four more goals with Dell in net and about five more goals with Jones in net.

Adding nine more goals against to their penalty kill tally would bring the Sharks’ kill rate from second-best to closer to league-worst.

For comparison’s sake, the following teams have all been good this season in terms of limiting shots, shots on goal, and scoring chances and/or expected goals against on the penalty kill: Pittsburgh Penguins, Arizona Coyotes, Vancouver Canucks, and Nashville Predators.

It would take a lot of video watching to understand whether this is a systems issue that needs fixing or if simply swapping players out would do the trick. A few changes that might help, based on how well the team does at limiting shots, shots on goal, and high-danger scoring chances against relative to the rest of the team when a given player is on the ice:

• Less Joel Ward and Chris Tierney; more Timo Meier, Jannik Hansen, Tomas Hertl, and Mikkel Boedker.
• More Joakim Ryan at the expense of Marc-Edouard Vlasic. Vlasic isn’t particularly great at penalty killing, and removing him from special teams would allow him to focus his energy on shutting opponents down at 5v5. /

Whether the team needs a systems fix or a player swap, they’d better figure out what ails their penalty kill soon, or — oh, there goes gravity — the goalies are going to start falling back to earth and the empty penalty kill won’t be there to save them.