Stat of the Union: A goaltending roundtable

Updated: April 10, 2017 at 9:27 am by Adam Laskaris

With a few off days before the NHL playoffs start, we thought we’d give you a moment to check out the latest edition of Stat of the Union: an NHL analytics roundtable. Today, we’re bringing in five of the best goalie enthusiasts out there to give you their takes, hot or not, on the current state of affairs with the last line of defence in today’s NHL. On today’s panel:

Cat Silverman (CS) 

I work for InGoal Magazine and FanRag Sports covering hockey with a specialty in goaltending. I used to coach with the Arizona Coyotes youth development program, before having my daughter this past September.

Ian Fleming (IF)

I am a writer for and with a focus on the statistical evaluation of goaltenders, a business analyst for an international, tech-centric legal services provider, and a former goaltender.

Emmanuel Perry (MP)

Creator of Corsica.Hockey and analyst for WinnersView.

Greg Balloch (GB)

Hey, I’m Greg Balloch! I am a writer for InGoal Magazine based out of Vancouver, BC. I also work as a private goaltending instructor, and with the Surrey Eagles of the BCHL as their video coach and goaltending consultant. I originally grew up in Hamilton, Ontario (where I played), but moved to BC in 2011 to cover the Vancouver Canucks for one of the local radio stations. I’m 26 years old, and spend most of my day on Twitter attempting to explain why goalies do what they do.

Nick Mercandte (NM)

I am an attorney and analyst/consultant. I have over 30 years of experience playing goalie and coaching goaltenders. I currently write for Hockey Graphs and have written in various other places in the past.

What was your introduction to getting involved in goaltending? Consider time, place, and initial feelings.

CS: I’m not a super great skater, so I initially chose to become a goaltender to mask my footwork deficiencies. Of course, the position has undergone a radical transformation since I started playing – so I do private skating work now in order to be a more effective coach.

IF: I became more involved in goaltending when I was made aware a couple of years ago that there were more statistics for goaltenders than all-situations save percentage, goals against average, and wins. It was exciting to see progress in statistical evaluation and I wanted to be an advocate for that progress.

MP: My involvement in goaltender analytics is incidental to my broader interest in general hockey statistics. As far back as I can remember, people have wished for more when it came to evaluative metrics for goalies. As the webmaster of one of the biggest hockey stats sites out there, I feel it’s my responsibility to provide innovative and interesting statistics that people may use to analyze goalies. If nothing else, I hope that with my public xG data, danger zones, GSAA and K ratings, I’ve given people a sturdy enough foundation to build upon.

GB: I was practically born with a goal stick in my hands. My father is also a goaltender/instructor, and while he never pressured me to become one, I always wanted to be like him. I started skating at 4 years old, spent the first 2 and half years learning the basics, then immediately switched to goalie. I never looked back! My whole family was involved in hockey. My mother was the head coach of my teams through the early years, as well as the president of Oakville’s girl’s hockey association. My 2 sisters also both played, but neither of them decided to play goal. So as you can see, my life has revolved around hockey since the very beginning. I’m extremely pleased to be in a position where I am constantly surrounded by the sport I love, and I continue to learn new things every day.

NM: When I was playing in the local mite youth program our coach asked if anyone was interested in strapping on the pads. I raised my hand without thinking. I then played both goalie and skater for several years as a young player (a path I highly recommend for any goalie) before finally switching to goalie full-time. The rest is history.

Who’s your favourite goalie to watch, and why?

CS: Carey Price is the obvious answer, but it’s really true. He just does everything with almost a mechanical precision, yet makes it look so smooth; I don’t think I’ve ever seen a goaltender in the history of the position just transition and set up as seamlessly and confidently as Price does. He’s a generational talent to watch.

IF: Henrik Lundqvist, hands down. Consistency is a nebulous term in goaltending and is often derived from suboptimal, traditional metrics, but the level at which Lundqvist has played for the entirety of his career is nothing short of generational. He plays at a higher level, and more frequently so, than any goaltender of this era.

MP: Goaltending is a very intricate art and I don’t pretend to understand more than a fraction of what goes into NHL-caliber netminding. That being said, it’s hard not to appreciate how easy Carey Price makes it look. Sometimes you forget that stopping a three inch rubber disk travelling at 100 kilometres per hour from entering a 4×6 cage is supposed to be difficult.

GB: I wouldn’t be surprised if every person answers this question with Carey Price. He is the prototypical “hockey school” goaltender. He does everything exceptionally well. Whenever I need a clip to show a goaltender how to properly execute a move or technique, Price is the goalie that I immediately look to. Just when you think you have him nailed down, you watch another game and he finds a way to surprise you. There are many great goaltenders in the NHL right now, but none can match the presence that he brings to the crease. He is an excellent ambassador for the game, and an absolute pleasure to watch.

NM: Henrik Lundqvist. His approach is cerebral and measured. Given how conservative he is in his crease, everything he does appears to require precise calculation. But it’s happening at the speed of the game. I appreciate precision. I also love watching Tuukka Rask. I played and coach to the Finnish style – the school and philosophy of Urpo Ylonen. Rask in his prime is a beautiful example of that style.

What’s one thing the average fan doesn’t understand about goalies?

CS: So few fans understand that the better the save, the less likely you are to notice it. The ‘wow’ factor in a lot of saves for non-goaltenders comes from desperation and poor positioning; the ‘wow’ factor for goaltenders and in-depth hockey analysts (including coaches, scouts, and other players) comes from the saves that were executed with simple perfection. If there’s a giant leap involved, there’s a good chance it wasn’t a technically excellent save. If a goaltender makes a save look easy through traffic? That’s talent.

IF: From an on-ice perspective, one thing that is easily lost is how active goaltenders are at all times. There are constant small adjustments in positioning to maintain angles and tracking of the puck, which lead to the big saves that our eyes are more attracted to. From a statistical perspective, what can be tough to understand is how heavily team effects and shot quality play into traditional metrics. Not all .910 save percentages were created equal and the value of that number is highly dependent on the quality of the shots against.

MP: I think this extends to the rest of hockey (or even sports generally) but understanding how pervasive luck is can be a hurdle for people. At this point, I think there’s ample evidence that the skill component in preventing low-danger goals is minuscule compared to plain variance. It’s funny that you have a good chance of hearing “if he sees it, he stops it” on any given broadcast, yet surprisingly few people understand that this lofty lower limit of talent means goalies can only really separate themselves by stopping more or fewer high-danger shots. On 2,000 shots, 2 fluke goals equates to a 0.1% difference in save percentage – the difference between second (Craig Anderson) and third (Devan Dubnyk) place in the league at the moment. Fluke goals happen just about every day in the modern NHL.

GB: The physical toll that the game has on your body. There’s a reason NHL goaltenders max out at 60-70 games a season today. It is an absolutely punishing position on your body, and that’s the first thing I tell people that want to start goaltending later in life. Your joints suffer, specifically. While a lot of older goaltenders complain about their knees, butterfly-era goaltenders have seen an increase in hip damage over the years. It’s rare these days to find an NHL goaltender that hasn’t had some kind of operation on their hip(s). Stretching and proper self-care is essential for survival if you’re a goaltender.

NM: Reps. Many fans don’t realize how technical the position is. And how much repetitive practice and attention to detail you need in order to play without thinking about it. Playing with perfect fundamentals and form for your body type and athletic style is a difficult balancing act that requires constant work and maintenance. That’s one reason why, in comparison to skaters, goalies on average take a bit longer to hit their performance prime. You’re still maturing and improving as a player. And maybe finally getting the quality of coaching to help you do so, which is harder to find as a youth.

How does your career experience (both inside and outside of hockey) give you a unique perspective that others may not have in relation to goaltending?

CS: I’ve gotten to work in both the media and on-ice side of hockey, so I’ve gotten to see both how the position has evolved from an in-game perspective and how the rest of the game has evolved with it. I think that a lot of journalists don’t understand the practical application of a lot of the elements of goaltending, but I’ve gotten to gain a much better understanding (and awareness) of the rest of the game by stepping back from the crease and writing about forwards, defensemen, and coaches as well.

MP: My expertise is statistics, data science, programming. I don’t know the first thing about playing goal, and I like to think that can be used to the benefit of my analysis. Unless you happen to have a near-superhuman memory, I think relying on observation to evaluate goaltenders can be a bit of a fool’s errand. The manner in which you stop pucks is remotely secondary to your plain ability to do so. I prefer to ignore the former almost entirely. Others who have an attachment to the position have a more difficult time doing so, often to the detriment of their ability to objectively evaluate talent.

GB: As a coach, I definitely look at the game through a very technical scope – but the journalist side of my job allows me to talk to people from all sides of the game. Not only do I get the pleasure of talking to some of the most talented and respected coaches/goaltenders in the game, but I also love to hear from the analytical side. The only way you can expand your knowledge about any subject is by keeping an open mind and hearing other opinions – and that includes hearing the opinions of non-goalies. If I can bridge the gap between on-ice technical analysis and statistical analysis even just a little bit, it will help me understand the game better and make me a better coach.

NM: Combining the analytical thinking and problem-solving skills of legal practice with my passion for hockey came naturally. I’m an attorney. I’m a hockey-obsessed hockey player. Slam the 2 things together and you get someone who isn’t willing to accept information at face value without vetting it and digging deeper as necessary.

What do you think is the best predictor of future success?

CS: With goaltending, there’s such a complex group of factors that go into successful execution – so that’s tough. I hate to say it, though, but coaching is my best indicator of success. Does a goaltender seem to be gaining more confidence in net under his coach? Is he being pulled out to his detriment? Is he being encouraged to incorporate outdated techniques, or is he being slowly introduced to newer and better tools? There’s a reason that everything Mitch Korn touches turns to gold – and that other teams in the league see great goaltenders arrive there to die. A good or bad coach can tell you a lot about how successful a goaltender is going to be; while his stats (shot distance, save percentage given outlying factors, danger save percentage, goals saved above average) can give you some good clues, a lot can change when a player arrives with a new boss.

IF: The million dollar question. Right now, there isn’t any one stat that you can hang your hat on as being repeatable and predictive. I believe it’s going to take an aggregation of metrics that can be attributed to a goaltender without team effects, weighted to each goaltender’s specific talent set, before we get reasonably close to predicting future success. We are close to getting to that point with the help of machine learning.

MP: This is a trickier question than you might first think. Some would think future success should be measured by save percentage or wins. Increasingly, we’ve tried to separate team effects from evaluative measures of goaltender ability, but if we can’t agree on what we’re trying to predict, some might find that their inclusion improves predictive validity. If the problem is how to best forecast pure goalie performance (as measured by GSAA, for instance), then I would say GSAA or FSv% relative to expected.

GB: Goaltenders exist in such a weird little spot. They have little-to-no control over the game around them. Everything that happens is the result of another action that is completely out of their control. It’s unlike any other position. This makes it immensely frustrating for analysts and coaches, because goalie statistics can wildly fluctuate from year to year. Understanding context is important to understanding why a goaltender is having success/failing – but it’s difficult at any level to fully grasp. Save percentage (and its various adjusted/reduced forms) is all that we really have to go off right now as a base stat. As a coach, I can identify warning signs and red flags in a goaltender – but incomplete goaltenders can still find success in the right system. Outliers will always exist because of this reason. There is no “money stat” that we can point to that can predict goalies. We’re better off than we were when goals-against-average and wins were the normal goaltending stats – but we still have a long way to go. Understanding which defensive systems work with certain types of goaltenders would be an excellent place to start.

NM: For goalies I think it’s 5v5 distribution of quality of performances within a controlled sample. The Win Threshold % statistic I discussed last year at RITHAC looks at what percentage of qualifying appearances a goaltender exceeded the 5v5 xGSAA/23 performance of an average winning effort (1 5v5 goal allowed). So it is looking at how often your goalie has hit a high ceiling performance (regardless of actual goals allowed, and in consideration of the types of shots faced. On the other side, Loss Threshold % looks at how often your goalie gave a “losing effort,” or performed equal or worse than a goalie giving up 3 5v5 goals (again, regardless of how many goals were actually given up, and in consideration of quality of shots faced). So it’s looking at consistency (doing enough to not sink your team). It is a lot to take in, I know. I should write about it, eh? The bottom line is this: each stat is mildly predictive. So we can start looking at chunks of performances, a distributions, and “what is this goalie generally doing.” We should stop looking at trying to predict sv%. It is getting us nowhere in the absence of better shot data.

How much do you hate GAA and goalie wins as evaluators of talent, and what do you suggest instead?

CS: Goals-against average can be a very, very mild indicator of success, and so can wins. I get a lot of flack for appreciating wins – after all, it’s very true that goaltenders don’t score goals, so they don’t actually win the game when all is said and done – but as long as you use those stats without looking at them in a vacuum (look at how many shots the goaltender is facing, look at shot distance, look at wins in comparison to other goaltenders on that team) you can get at least some idea of how things are going. Is a goaltender winning sixty percent of his games, but only facing 20 shots a night? He’s possibly not as effective as the goaltender who’s winning sixty percent of his games facing 35 shots a night, and from ten feet closer in.
If I had to choose a stat that I think best evaluates talent…. hmmmm. Here’s my hot take: no one goaltending stat gives you a true evaluation of his talent. You need to look at them all, you need to look at them over time (consistency is key), and you need to look at how they compare to other goaltenders in similar situations.
Sorry I’m not more help.

IF: GAA and wins are team stats, should be treated as such, and do not belong anywhere in the conversation of talent evaluation. Right now, in my belief, the best places to look for statistical evaluation are those metrics which incorporate shot quality/expected goals. With an indicator of how many goals should be scored, we can adjust and evaluate for how a goaltender has performed in relation.

MP: Quite plainly, if you use GAA you don’t know a thing about goaltender evaluation.

GB: I’ve written about this many times on Twitter, and at length in InGoal Magazine. GAA is a team stat. I call it “[TEAM]’s goals-against-average when ____________ is in goal,” because that’s exactly what it is. It’s about as valuable as catcher’s ERA in baseball. Wins are important from a team perspective, but have no place when individually analyzing or comparing two goaltenders. Goals saved above average (or GSAA) is the best hard statistic that we have today when evaluating the type of season a goaltender is having. It simply takes their save percentage, compares it to the league average goaltender, and tells you how much the goalie has helped or hurt their team. Breaking it down even further with 5-on-5 adjusted save percentage rather than raw save percentage is useful as well.

NM: Goalies can’t score. Teams give up goals against. Not just goalies. Don’t make me angry.
What’s your view on the low-scoring debate in hockey? Do you have any ideas on how to increase (or decrease) scoring?
CS: Honestly? I think the game is fine just the way it is. I’m sick of all the complaints. We’re seeing some of the most talented hockey fans have ever seen; if they want more high-scoring games (which in previous generations came from player execution errors more than anything else) they should go ahead and buy season tickets for an SPHL team.

MP: I desperately want to see more goals, but more specifically, I desperately want to see more 5-on-5 goals. I’ve seen plenty of suggestions for increased scoring that involve more more power play time and I really don’t think this addresses the issue. It does, if you think the issue is the NHL lacks excitement and that goals are exciting. But my concern is one of false parity and randomness. There are too few goals, particularly at full-strength, for teams to separate themselves in terms of true ability. Shooting and save percentage normalize much quicker if goals are more frequent, and 5-on-5 play is the truest representation of a team’s quality. Make the nets bigger and the better team is more likely to win on any given night.

GB: I get my kicks out of watching goaltenders go to work, but for the more complete fans of the game, I can understand the frustration at the lack of scoring in recent years. Tinkering the game with new faceoff rules, and new gear restrictions is not going to help in any noticeable way. What players need is more time and space. As we see when it goes to 3-on-3 overtime, these are the best shooters in the world. If you give them room, they will score on any goaltender of any size. NHL games are so overcoached, this rarely occurs when it is 5-on-5. Nobody wants to be creative, because creativity sometimes leads to mistakes, and mistakes lead to goals against. The teams that shut the game down defensively the best will always have the best chance at winning. I remember tracking playoff games from 2 years ago, and there were games when Ben Bishop (playing with a pulled groin) had literally 2 legitimate scoring chances on his net the entire game. This is not an issue with the goaltending. It is systematic.

NMTiny pucks! Really I’m cool with the way the game is.

Do you have anything else to add?

CS: Goaltending is voodoo. There. I said it.

IF: Keep your eyes on the goaltending space of analytics in the near future. This is the area of the sport that has the greatest potential for growth and there are a few things that I and others I’ve spoken with are working on that, I think, are very exciting.

MP: Just Win Baby.

GB: I love talking about goaltending! As you can see, it is quite literally my life. I’m always down for a discussion on Twitter, and you can find me at @GregBalloch. If there’s a question you need answered, or just want to gush about a big save that Carey Price made, hit me up!

NM: Follow me on twitter @nmercad and check out Hockey Graphs for cool goalie stuff this summer!