Anyone who has spent any amount of time within the comments sections of hockey articles or within the hockey circles of Facebook, Reddit or Twitter is likely well aware of debates that constantly occur regarding the validity and importance of advanced metrics. These arguments range in complexity and ferocity, but the vast majority of time they involve metrics used to evaluate the play of skaters or teams as a whole. However, the advanced stat-related arguments that permeate blogs and social media rarely venture into the realm of goaltender play analysis, despite there being plentiful debate about goalie play. Ranger fans in particular have been subject to a myriad of arguments regarding goalie play the past few years, as Henrik Lundqvist is no longer at the peak of his hall of fame-caliber powers, and the defense in front of him has a proclivity for allowing high danger chances.
Recently, a couple friends of mine reached out asking if there was even such a thing as “advanced stats for goalies,” and it dawned on me that a likely reason for why goaltender analysis conversations include advanced stats much less frequently than skater and team-related debates, is because fewer people are aware they even exist. I reached out to a few of the Patreon subscribers for the Blueshirts Breakaway podcast and asked if there was an appetite for some discussion and explanation of goalie advanced stats, and well, considering you are currently reading this article, you can guess how they answered.
In this article, I will summarize some of the most useful advanced goaltender metrics available to the public, including where you can locate them and how to use and interpret them. I will also recommend some of my personal favorite hockey analysts who have taught me a lot about goalie play and analysis.
What are the Advanced Goalie Stats and How Do I Use Them?
Currently, goaltender discussions are at the level that baseball discussions were 10 years ago, where many fans over-emphasize a team success stat like wins to evaluate a goalie’s performance and rely heavily on metrics that can be significantly impacted by the play of your defense. In the case of goalies, those metrics are simple save percentage and goals against average, whereas for pitchers its ERA. All are flawed statistics that can be influenced by the play of the defense and only offer a specific perspective on how the player performed (yes, I know ERA doesn’t include runs scored as a result of an error, but the quality of your defense along with a litany of other factors absolutely plays a role here).
Now, I’m not saying that numbers such as save percentage for goaltenders and ERA for pitchers aren’t helpful, they certainly are. What I am saying, however, is that they are flawed, and there are other metrics out there that we should also be considering that help isolate the performance of the actual player, and strip out some of the noise in the background that may be influencing the more basic numbers. So, what are these additional goaltender statistics, and how can you use them to evaluate goalie performance? Here are the ones that I personally find most helpful, along with examples of how you can use them in your own analysis going forward.
Most fans I see on social media, regardless of their feelings about advanced stats, seem to agree that shot quality certainly matters, and will use shot quality principles when examining whether a goal was a goalie’s “fault” or not. With that in mind, a handful of hockey stat sites, including my personal favorite Corsica, break down goaltender save percentage into individual save percentages by shot quality. A number of factors go into determining the “quality” of a shot, such as the area of the ice a shot is taken from, the type of shot, the angle of the shot etc. For a full breakdown on exactly how Corsica’s model works for categorizing shots by quality, and therefore determines save percentage by shot quality, you can read Emmanuel (Manny) Perry’s Corsica blog post about it. Corsica provides the following shot quality save percentages:
- Low Danger Save Percentage (LDSv%) – Save percentage on shots with a Fenwick shooting percentage of 3% or under. Fenwick shooting percentage is the shooting percentage on all unblocked shot attempts, and not just shots on goal. This serves as a far more accurate representation of a player’s actual shooting percentage, compared to the standard shooting percentage number which only accounts for shots on goal and ignores all of the times the player shot the puck and missed the net.
- Medium Danger Save Percentage (MDSv%) – Save percentage on shots with a Fenwick shooting percentage equal to or greater than 3%, but less than 9%.
- High Danger Save Percentage (HDSv%) – Save percentage on shot with a Fenwick shooting percentage equal to or greater than 9%.
One thing that Manny noted in his piece that I think is worth explicitly stating here, is that in large sample sizes, most NHL-caliber goalies put up similar low and medium danger save percentages. Given this, “it appears the skill-driven component of Sv% is almost entirely contained in a goalie’s ability to stop shots of the High-Danger variety.”
These shot-quality save percentage buckets, and the expected goals data that serves as the foundation for them, help provide the following metrics that attempt to strip out the impact that defenses have on a goalie’s save percentage, and paint a more accurate picture of a goalie’s actual performance:
- Expected Save Percentage (xSv%) – The save percentage that a goalie should have with a league average performance given the quality of chances he faced. It is important to note that expected save percentage is NOT a measure of how the goalie has actually performed, it merely serves as a benchmark for the save percentage that an average goalie should have posted, given the quality of chances he faced. Given that, please don’t ever be that guy to make an argument I recently saw on Twitter, where a guy said that because Halak’s xSv% is greater than Lundqvist’s, that means Halak has had a better season. That is a very false statement. What that does mean, however, is that the Islanders defense has performed better in front of Halak than Lundqvist’s has, so Halak has faced on average lesser quality scoring chances than Lundqvist has.
- Delta/Adjusted Save Percentage (dSv%) – This stat is the difference between a goalie’ expected save percentage and his actual save percentage. This is a very valuable stat that helps show how much better (or worse) a goalie is doing compared to how an average goalie would have performed given the quality of shots he has faced. A dSv% of 0 means that a goalie has performed exactly to the level of an average goalie given the quality of shots he has faced. Using our Lundqvist and Halak example from before, Lundqvist hasn’t been great this year (although he’s been much better of late), and is currently sporting a dSv% of -0.22 during 5v5 play, which is 18th in the NHL among goalies with at least 300 minutes played. However, it is considerably better than Halak, who has a dSv% of -0.94 during 5v5 play.
- Goals Saved Above Average (GSAA) – This is a cumulative stat that represents the number of goals allowed by a goaltender compared to the number of goals that would have been allowed by a league average goalie. It is similar in nature to WAR in baseball that way, except it is specific to goals allowed by goaltenders. Currently, Sergei Bobrovsky leads the NHL with a GSAA of 8.87 in 5v5 play, meaning that, if an average goalie was substituted in for Bobrovsky for all of his time played this year, and the games unfolded exactly the same way, that the Blue Jackets would have yielded about 9 more goals on the season in 5v5 play. Columbus is currently tied for 13th in the NHL with 50 goals allowed, so according to this stat, if you substituted an average goalie for Bobrovsky this year, the Blue Jackets would have allowed about 59 goals, which would drop them all the way down to 25th in the NHL, tied with the Capitals and one behind the Rangers.
- Goals Saved Above Average per 30 Shots (GSAA30) – As I’ve discussed multiple times in previous articles, rate statistics help us level the playing field and compare players who have had unequal playing time. GSAA30 is the rate version of GSAA, and tells us how many goals saved above average a goalie has per every 30 shots he has faced. Cory Schneider currently leads all NHL goalies in GSAA30 in 5v5 play (minimum 300 minutes played) with 0.78, and Bobrovsky is second at 0.73.
Where Can I Access Them?
I’ve given you links above to Corsica, which is the site I personally use most often for hockey data. However, I have had a number of people ask me how to use the site to find what they are looking for, and fans that don’t geek out about stats the way I do have often said that they find the site to be overwhelming. My job in real life is actually that of a user experience analyst for websites, so I can tell you that I definitely can understand where fans are coming from, and I know from my own professional research that the quickest way to piss of a user and make them not want to come back to your site is to make them feel confused or overwhelmed.
With that said, I think Corsica is a valuable resource, despite some of its user experience flaws, so let me give you a quick demo about how to use it to access the advanced goalie stats, in the hopes that you will begin to learn how to use the site and enjoy all that it has to offer.
When you arrive on Corsica’s homepage, you can access the goalie stats from the Goalies dropdown at the top, or the Goalie Stats image link in the center of the page.
Selecting either of those Goalie Stats links leads to the Goalie Stats Summary Report page, which by default lists applicable players within the table below in alphabetical order. The page by default will be set to include players from all teams but only includes data from 5v5 play. The top of the screen has a number of data filter options, including the ability to filter the data by any game state (all situations, 5v5, PP, PK etc.) and by season, team, venue and session. You can also place a time on ice minimum for the players you wish to include, and you can search for a specific player. Further, the table is fully sortable, so you can sort it by any of the columns by simply clicking the column header. Lastly, you can choose to view more detailed reports (reminder, you arrive on the Summary Report) from the Report dropdown at the top.
The Summary Report contains most of the stats I discussed in the previous section, as well as additional details such as games played, total time on the ice, shots against, goals against and standard save percentage. The left side of the screen provides the standard statistics, while the six right-most rows display the advanced stats we discussed: xSv%, dSv%, LDSv%, MDSv%, HDSv% and GSAA. However, if you wish to view the GSAA30 data, you must use the Report dropdown at the top and select the Shots Faced Report, which includes a litany of raw data totals that feed into the data on the Summary Report, such as expected goals against, shot attempts per danger zone and Fenwick against. Depending on the size of your browser, you may need to scroll to the right to find the GSAA30 column, which is the one furthest to the right.
Where Can I Learn More?
There are a number of excellent analysts on Twitter that are worthy of following if you wish to learn more. To be honest, I feel bad even including this section, because I know for a fact that I am leaving off a number of fantastic people. However, here is a list of six analysts that I personally follow, all of whom have been great to interact with and have helped me learn a lot about goalie analysis. If you have any suggestions for analysts that I did not include here, please feel free to leave a comment naming the analyst, or shoot me a message on Twitter.
- Emmanuel Perry – Creator and proprietor of Corsica and overall excellent hockey analyst who has created multiple, extremely useful analytical models that many analysts use on a daily basis.
- Nick Mercadante – Easily among the most notable goalie analysts in the industry, and contributor to HockeyGraphs and frequent guest on Dimitri Filipovic’s PDOcast (I highly recommend this podcast if you don’t already listen, and I also recommend following Dimitri’s work). He is a former goalie himself, and is also a New York Rangers fan.
- Garret Hohl – Co-Founder of HockeyData, which is a leading provider of hockey analytics and information to teams, players, agents and scouts. Frequent guest on the PDOcast, and also manager (and I believe co-founder) of HockeyGraphs.
- Ian Fleming – Contributor for HockeyGraphs and NHL Numbers, and the creator of the excellent hockey data visualization tool, SAVE, which functions similarly to Domenic Galamini’s HERO Charts.
- Cole Anderson – Founder of crowdscoutsports.com, and creator of the excellent Goalie Compare App. He is also easily among the most pleasant analysts to interact with on Twitter (to be clear, everyone on this list is fantastic to interact with, but Cole has always been great with answering any questions I have had).
- Steve Valiquette – Former Rangers goalie and current studio analyst for the Rangers (which most of you probably already know). He is also CEO of the hockey data company Clear Sight Analytics, and easily one of my personal favorite analysts across any sport.
Cat and Dog Goalies
I will leave you all with this, some wonderful gifs of dogs and cats making saves (courtesy of Hayley “Elfie” Nath, Patreon subscriber to the Blueshirts Breakaway):
Author: Drew Way
Diehard New York Rangers fan since 1988! Always has been fascinated by sports statistics, and is a big proponent of supplementing analytics with the eye test. Also a big Yankees, Giants and Knicks fan.