Stop Protecting Your Goalie

Three Slots

Phase 1 of the PWHL's expansion process gives each existing team exactly three protection slots. Three. That's it. Every player you don't protect is available to Detroit, Hamilton, Las Vegas, and San Jose through the Exclusive Negotiation Target Lists in Phase 2, or outright selection in Phase 4.

You get three more protections in Phase 3, but by then the expansion teams have already had a week to sign your unprotected players. The first three slots are the ones that matter.

So when Montreal protects Ann-Renée Desbiens, Boston protects Aerin Frankel, Ottawa protects Gwyneth Philips, Vancouver protects Emerance Maschmeyer, and Minnesota leans toward protecting Maddie Rooney, a question: are these goalies really more irreplaceable than the skaters those teams are leaving exposed?

The Supply Problem That Isn't

The conventional wisdom is that expansion creates a goalie shortage. The Hockey News wrote that goaltending is “the only position that likely cannot be filled primarily through the draft.” Coverage has treated the goalie market like a crisis.

But let's count. Four expansion teams need four starting goalies. Here's who's available:

That's six proven PWHL goalies plus two elite draft-eligible prospects for four open starting jobs. There are more qualified goalies than there are teams that need one. This is not a crisis. This is a buyer's market.

Save Percentage Is a Team Stat

Before we even get to the leaderboard, we need to talk about what save percentage actually measures. The short answer: not just the goalie.

A goalie's save percentage is heavily influenced by the defense playing in front of her. Better defensive teams allow fewer shots, and more importantly, fewer high-quality shots. A goalie facing 20 shots from the perimeter will post a higher SV% than a goalie facing 20 shots from the slot, regardless of individual ability. The NHL developed “expected save percentage” (xSV%) and “Goals Saved Above Expected” (GSAx) specifically because raw SV% is misleading. The PWHL doesn't publish shot quality data, so we can't isolate how much of any goalie's numbers are her and how much are her teammates.

NHL research has quantified this. Tom Awad found that only about 30% of a goalie's deviation from league average actually belongs to the goalie. The rest is shot quality, team defense, and noise. Separate research found that starter and backup save percentages on the same team correlate at roughly .58 after reliability adjustment, meaning team defense alone explains about a third of the long-run variance in save percentage.

The PWHL's own data makes this case beautifully. Look at what happens when backup goalies step in behind the league's best defenses:

Elaine Chuli, a backup, posted .949 behind Montreal's defense in Season 1. That was the best save percentage in the league. Better than Desbiens. Klara Peslarova stepped in cold for Boston, posted .937 in 4 games, and threw a 29-save shutout. These are not elite goalies. These are backups who looked elite because they played behind elite defenses.

Meanwhile, Corinne Schroeder posted .919 as a starter on last-place New York. Good goaltending. The team still finished last. If goalies made teams good, New York wouldn't have been in last place.

The NHL has seen this over and over. Devan Dubnyk posted .916 in Arizona, got traded to Minnesota mid-season, and finished the year at .936, becoming a Vezina finalist. Same goalie, same season, 20-point jump. Robin Lehner went from .908 behind Buffalo to .930 behind the Islanders' defensive system. 22-point jump, Vezina finalist. Braden Holtby was elite behind Washington for years, then dropped to .905 after leaving. The numbers follow the team, not the goalie.

So when someone says Desbiens' .955 proves she's irreplaceable, ask: irreplaceable compared to her own backup, who posted .949 behind the same defense?

The Leaderboard Is Tighter Than It Looks

Here are the full 2025-26 PWHL goaltending stats for every goalie who played 12 or more games:

At first glance, Desbiens and Frankel look like they're in a different class. But look at the team context. Those two played behind the league's two best rosters. Kirk, who played behind a middling Toronto team, is the one who arguably had the most impressive individual season: .934 SV% with no defensive cushion. She's a pending free agent that Toronto may not even re-sign.

Strip out the team effects and the spread compresses further. The gap between Philips (3rd, .931) and Osborne (10th, .906) is 25 points. The gap between Desbiens (1st, .955) and Philips (3rd) is 24 points. The goalie everyone treats as third-best is closer to the worst starter in the league than to the best. If you adjust for shot quality, which we can't fully do without PWHL tracking data, the differences between starters 3 through 10 may be even smaller.

Goalies Don't Predict

Even if you believe your goalie is meaningfully better than the available alternatives right now, the data suggests you shouldn't bet on that holding.

NHL analytics research has consistently found that goalie performance is among the least predictable metrics in professional sports. The average NHL goalie experiences a year-to-year variance of .361 GSAx/60, which translates to roughly 13-14 goals of swing per season. That's about one goal every three games that you can't predict from last year's numbers.

The difference between an NHL starter and backup? About 0.006 in save percentage. Six saves per thousand shots. Over a 30-game PWHL season, that gap is worth maybe 3-4 goals total. A single power play or a single defensive breakdown erases that margin entirely.

This is the “outlier season trap” that NHL front offices have fallen into repeatedly. Philipp Grubauer posted a .922 SV% in Colorado, got a $5.9 million deal from Seattle, and fell to .889. Jack Campbell went .937 for a stretch in Toronto, signed a $25 million contract in Edmonton, and fell to .888. Carter Hart, Korpisalo, Matt Murray after Pittsburgh. The list of goalies who looked great for one window and then reverted is enormous. Every single one of them played behind a good team during their hot stretch.

Analysts put it bluntly: “Predicting goaltenders from year-to-year is still a nightmare.” If a goalie is above average in a given season, there is only a 59.2% chanceshe'll be above average the next year. Barely better than a coin flip. Less than 30% of elite goalie seasons are followed by another elite season.

And that's in the NHL, where goalies face 1,500-2,000 shots per season. The analytics community generally wants at least 1,500 shots before drawing any conclusions about a goalie's true talent level. Some set the threshold at 15,000. PWHL starters face roughly 600-700 shots per season. The data is so noisy that we genuinely cannot tell you how good any PWHL goalie is independent of her defense. We're making protection decisions based on a stat that is one-third goalie, two-thirds team, measured over a sample too small to separate the two.

The Opportunity Cost

This is the core of the argument. When you protect your goalie, you're not just saying “I want to keep this goalie.” You're saying “I would rather lose one of my top four skaters than lose this goalie.”

Every team has more than three skaters worth protecting. The Phase 1 list forces brutal choices. Protecting a goalie means leaving an elite forward or a top-pair defender exposed to four hungry expansion teams with exclusive negotiation rights.

Consider Vancouver. They protected Sarah Nurse, Sophie Jaques, and Maschmeyer. Maschmeyer posted a .916 SV% and a 7-9 record. Meanwhile, Kristen Campbell is right there at .913 and is available anyway. Vancouver burned a protection slot to keep a goalie who is statistically indistinguishable from the goalie they're willing to lose. Whatever skater they left unprotected to make that decision was almost certainly harder to replace.

Or consider the irony of Montreal's decision. They protected Desbiens, the goalie with the best save percentage in the league. But save percentage is partly a product of the defense. If Montreal loses a key defender because they used that slot on Desbiens, Desbiens' save percentage next year could drop. They may have protected the stat instead of the cause.

The math is simple: elite skaters are scarce. The PWHL has maybe 40-50 genuinely impactful skaters across 8 teams. You can find a .910-.920 goalie. You cannot find another top-line forward.

The Bottom Line

Four expansion teams need four starting goalies. There are at least six qualified options available through free agency and unprotected lists, plus two elite draft prospects. NHL data shows goalie performance is volatile and the gap between starters and backups is tiny. The primary metric everyone uses to evaluate goalies, save percentage, is partly a measure of the defense, not the goalie. And the PWHL season is 30 games long, which means even the “real” goalie performance signal is buried in noise.

Protect your skaters. All three slots. You'll find a goalie.