NHL Goaltender Volatility: How Goalies Replicate Elite Seasons

Goaltending is the hardest aspect of the NHL game to analyze and project accurately. Players' performance at this position is known to be so volatile that the phrase ‘Goalies are voodoo’ was coined to describe their erratic statistical production, and they are rarely picked at the top of the NHL Draft.

As a result, it has been argued that it’s never worth spending at the top of the market at the position because the men who mind the crease are unstable assets.

While that line of thinking is valid, it doesn’t seem to match how NHL teams are thinking right now. Six of the top nine active goaltending contracts by average annual salary were signed since July 1, 2023, when Ilya Sorokin earned an eight-year extension from the New York Islanders. 

On December 8, Igor Shesterkin reset the market with a $92 million — the largest contract ever for a player at his position. That contract, Sorokin's and Connor Hellebuyck’s (7 years, $59.5 million), represent particularly strong leaps of faith as the majority of each player’s expensive deal will be paid out during their thirties when production at the position has historically trailed off.

Other goaltenders with recent new deals were either younger, like Jeremy Swayman (25 at the time of his eight-year extension), signed for a shorter team like Linus Ullmark (four years), or made less than $8 million against the salary cap on their contract like Juuse Saros ($7.75 million).

Not only were the New York Rangers, New York Islanders, and Winnipeg Jets willing to take the plunge with their big-name goaltenders, but it appears public sentiment is behind those decisions. 

RG ran a Twitter poll to ask which one of the three goaltenders fans would want on their team, and as of publication, just 21.8% said none of them were worth their contracts.

The majority of respondents were most interested in Hellebuyck, whose career steadiness contrasts his position’s reputation for year-to-year inconsistently.

It seems that right now, both fans and teams are OK with the practice of paying top dollar to goaltenders. But are high-performing goaltenders immune to the swings in performance the rest of their peers seem to experience? 

That’s what this study aims to find out by looking at goaltenders who produce elite seasons and where their careers go from there.

If we started by trying to define an elite goaltender, that criteria alone would create a sample of exclusively consistent performers in the Hellebuyck mold that would defeat the purpose of asking how trustworthy they are. It would also shrink the sample to too small a group of players to learn much from.

So, we will look at goaltenders who had elite seasons and how they performed in the three years following. In other words, when a goaltender produces at an extremely high level, what are the chances he’ll keep up that level of play — or something close to it — in the near future?

Because we are looking at goalies at the top of their game, it’s fair to expect some regression to the mean. What’s important to consider is the extent. If a goaltender shows they can be one of the best in the league in their position, are they players that teams can safely invest in?

The Criteria for this Research

Let’s start with a working definition of an elite season. Here are the criteria we’re using.

• At least 41 games played: Anything less than half an entire season is too small a sample to be considered here. In shortened seasons, this limits the pool, but bringing the standard for games played could invite some goaltenders who aren’t even the primary starter on their own team into the sample, which doesn’t fit the spirit of this research.
• A save percentage (SV%) in the top 10 of NHL goaltenders with at least 41 games played: Save percentage (which is calculated by dividing saves by shots against) is arguably the single most-used statistic for evaluating the play of goaltenders. It answers the question, “How often does this goaltender let pucks in?” which gets to the heart of goalie play.
• A GSAA in the top 10 of NHL goaltenders: The statistic from Natural Stat Trick measures “The difference between the goalie's goals against and goals against with the same Shots Against and the average SV%.” That might seem like a subtle difference from SV%, but it uses a stat called expected goals to determine how an average goaltender would perform against all shots based on location. That makes it slightly more sophistical than SV% because goaltenders are credited more for stopping shots that are more likely to score. This measure should help weed out goaltenders who produce a high SV% due to a strong defensive team in front of them, causing them to face less difficult shots. 

There are two cases where goaltenders that meet these standards are weeded out:

1. They play less than three seasons after their elite season: The performance metrics after the ‘elite seasons’ needs to be uniform for all players in the sample.
2. They are over the age of 35 in their ‘elite season.’ At this point in a player’s career, a significant decline is expected. If a goaltender performs poorly from age 36 and beyond, that probably says more about physical wear and tear than overall goaltender volatility.

With this definition in mind, let’s take a look at a decade of impressive NHL goaltending seasons from 2011-12 to 2020-21 (ending at a point where there are three full seasons afterward to evaluate) and see what we learn about how reliable the top performers are.

In the first three years, relatively few seasons by goaltenders meet our criteria (14), and there are fewer goaltenders showing up in multiple years than we see later on. 

The save percentage numbers these goalies produce may look high compared to active players, but it’s worth remembering that the average SV% has declined significantly in recent years.

That context is helpful for understanding the goaltenders of the early-to-mid 2010s.

2011-12

• The decrease in the save percentage may look relatively modest, but where it went the three seasons after 2011-12 was right in line with league norms at the time. That meant the GSAA average dropped 99.7% as this group went from elite to almost exactly average.
• Lundqvist is the only one of these goaltenders to produce an elite season in the two years following 2012-13, though Quick shows up significantly later on (2017-18).
• Smith dropped off most of this group thanks partly to a rough 2014-15 that saw him lead the NHL in losses (42) and goals against (187).

2012-13

• This lockout-shortened season limits our sample, but the results still hew closely to 2011-12.
• Once again, Lundqvist is the only mainstay, and once again, the other goaltenders put in below-average performances by GSAA in the next three years, bringing the whole group to almost precisely average.

2013-14

• This is more what you’d expect to see as three of the goaltenders (Rask, Varlamov, and Price) put up elite seasons the next year, and another (Bishop) got back on the list two years after.
• Price stands out, going from an excellent 2013-14 to an even better stretch in the following years by SV%.
• The group’s GSAA average in the three years following their elite 2013-14 declined by 72.6%, but only Bernier was notably below average in those seasons.

This period has far more elite goaltending seasons than the previous three (24), and the same names are starting to come up a little more consistently. 

The trend of regression to the mean endures, but the goalies that produce standout campaigns during this time carry their success forward better than those in the previous three seasons.

2014-15

• These numbers closely mirror 2013-14, with a larger sample of goaltenders, with the average GSAA decline being 74.7%.
• While three of the nine goaltenders (Mason, Schneider, and Varlamov) end up below-average by GSAA over the next three seasons, none of them are by a significant margin.
• Three of the nine goalies produced an elite season in 2015-16 (Crawford, Schneider, Holtby), and two more (Price and Dubnyk) managed in 2016-17.

2015-16

• The save percentage decline is similar to other groups, but because the NHL’s average SV% began to drop after 2015-16, the performance of this cohort relative to its peers after their elite seasons is strong.
• GSAA drops off by 51.3%, which is the lowest rate of any group of goaltenders in this research over a full-length season.
• Only one of these goaltenders (Holtby) produced an elite season the following year, and just one more (Fleury) made the list the year after that despite the solid overall performance.
• This is the last time Crawford appears in this research, but he was a notable example of the kind of stability that is so rare in goaltenders during this time period. Between 2013-14 and 2016-17, he played at least 55 games per season, with a save percentage between .917 and .924 each season.

2016-17

• The results here are similar to the years above, where elite performance turns into approximately average results in the following years.
• What differentiates 2016-17 is that for the first time, we see two goaltenders with an average of -6.0 GSAA/season or worse in the following three seasons to help drag down the average.
• One of them (Talbot) struggled in the immediate aftermath of this season before bouncing back later in his career. Between 2019-20 and 2023-24, his save percentage was above average in four of five seasons, never dropping below .911 in those years.
• The other (Holtby) is a good cautionary tale for investing in elite goaltenders. Between 2011-12 and 2016-17, he ranked first among all goalie in shutouts (30) and fifth in save percentage among those with at least 250 games played (.921). Although he helped the Washington Capital win the Stanley Cup in 2017-18, from that season on, his save percentage was a below-average .905, and he was out of the NHL due to injury before his 33rd birthday.

As we get closer to the current day, there is an uptick in consistency, though the sample shrunk considerably in the last two years due to Covid-shortened seasons.

2017-18

• This group performed slightly better than some others in the aftermath of its elite season, but the vast majority of the heavy lifting was done by Hellebuyck and Vasilevskiy, who have combined to win the Vezina Trophy in three of the last six seasons. Unsurprisingly, they’ll pop up again.
• Quick’s career is fascinating because he’s a Conn Smythe Trophy winner who’s been an enormous part of two Los Angeles Kings championships, but he’s also been one of the worst goaltenders in the league at times in his career. Between 2018-19 and 2022-23, his save percentage dropped below .900 (.897), and his GSAA was -60.74. Much of that was attributed to age-related decline as he entered his mid-thirties, but in the last couple of seasons, he’s had a career renaissance with the New York Rangers as he approached his fortieth birthday. 

2018-19

• These numbers look awfully familiar as what constitutes an elite season remains relatively consistent even as league-wide save percentage declines, and most goaltenders who produce one revert to approximately average performance levels in the years afterward.
• Murray is another cautionary tale worth noting. By the end of his age-22 season, he’d helped the Pittsburgh Penguins win two Stanley Cups, and he was just 24 during the 2018-19 campaign. The wheels fell off quickly after that, though, and he’s currently fighting to remain in the NHL. On December 20, he made his first appearance in the league since 2022-23 after missing 2023-24 with injuries and starting 2024-25 in the AHL.

2019-20

• The COVID-19 pandemic shortened the season in a way that made only these two workhorses fit our criteria. There isn’t much to be made from the GSAA improvement because the following seasons had more games in them on average, but both goaltenders continued to succeed consistently after 2019-20.

2020-2021

• This time, Hellebuyck was slightly better than Vasilevskiy, both during the season in question and beyond, but otherwise, another shortened campaign yielded a similar result.

• Regression to the mean is a powerful predictor of goaltending results: We expected goaltenders to struggle to replicate single-season excellence over a three-season span, but the degree to which goalies who post elite seasons see their production decline towards league averages after is significant.
• Betting big on goaltending is risky: For every Hellebuyck and Vasilevskiy, there is a Murray, Holtby, or Quick who can’t be counted on soon after producing excellent results. Price was highlighted in this research for his exceptional statistical accomplishments, but he played his last NHL game before turning 35 with four years remaining on his contract.
• Showing you can perform at an elite level is still a good indicator for longer-term success: While many of the goaltenders in this research were more average than exceptional after one incredible year, it’s notable that over 60% of elite seasons were followed by three-year spans of above-average goaltending according to GSAA.
• Even as save percentages fall, the top end of the performance curve is relatively stable: In full seasons in this research, the average save percentage of a goaltender having an elite season remained between .923 and .927 in any given year.