Penalty Kill Save Percentage - PKSV%

Introduction

In a recent article, we introduced the metric known as the even-strength shutout (EVSHO). The idea for this metric stemmed largely from a desire for consistency in goalie analysis. As a motivator for the EVSHO, we used data that revealed that a goalie's save percentage when his team is on the penalty kill (PKSV%) was non-repeatable. That is, a goalie who posts strong PKSV% numbers one year is unlikely to post strong numbers in the subsequent year.

This non-repeatability is important if your aim is to project a goalie's save percentage in the upcoming season. This has obvious implications both from a hockey and fantasy hockey perspective. For example, imagine you are the Vancouver Canucks and you're looking to trade Roberto Luongo and hand the starting goalie reigns over to Cory Schneider. Is this a good decision? Well, there are a number of avenues of thought you should explore in a decision of this magnitude. One of these avenues is to consider the PKSV%. Schneider posted a total SV% of 0.937 in the 2011-2012 season where he played in 33 games (28 starts). This SV% was good for 2nd best in the NHL last season. Cory Schneider's PKSV% was 0.959 - a league best. Before you go off running down Robson street yelling for the Cup, you should know that an important detail lurks beneath the surface: Schneider's overall SV% was inflated by numbers that he doesn't control and numbers he likely cannot sustain (his PKSV%).

Shots Against: The Great Equalizer

Since the lockout, no goalie who has faced at least 400 shots against on the penalty kill (PKSA) has been able to maintain a PKSV% better than .892. Sometimes seeing data is more powerful than reading it. Below is a graph with data for all active NHL goalies where we've plotted PKSA (shots against while on the penalty kill) versus PKSV% (a goalie's save percentage while on the penalty kill).

For a better handle on the vertical axis, consider that 400 PKSA is roughly equivalent to a full season (82 games). Additionally, the post-lockout league average for PKSV% is .875 annually with little variance.

What this plot tells us is that the more pucks you face, the more likely your PKSV% is to move below .892. To date, Cory Schneider (2nd dot from the right) has faced 358 PKSA in his career and is holding onto a .925 PKSV% over the course of those shots. So, if you're Vancouver management or a fantasy hockey manager who owns Cory Schneider, there are two ways to interpret the data:

1. There is a measurable and distinguishable talent-level for goalies in non-EV situations AND Cory Schneider is the best goalie in the league in these situations.
2. Cory Schneider has benefitted from small sample sizes and his PKSV% (and overall SV%) are headed downward.

Which camp are you in?