Why Devils’ Unorthodox Goaltending Plan Makes Sense
The New Jersey Devils entered the offseason with a pretty obvious problem in net.
Jacob Markstrom had just completed a disappointing second season in New Jersey, and his $6 million annual cap hit was not looking particularly appealing. The Devils needed to find a way to improve their goaltending while simultaneously improving the depth of their roster — which certainly wouldn’t be easy.
General manager Sunny Mehta did a great job given the circumstances. He traded Markstrom to the Florida Panthers, brought Evan Rodrigues to New Jersey and created significant financial flexibility in removing Markstrom’s entire $6 million cap hit.
A few days later, they added David Rittich to join Jake Allen and Nico Daws in the crease. None of those goalies are viewed as a star. None are entering the season as an unquestioned number one.
For some Devils fans, that’s concerning. They’re holding out hope that more of a surefire No. 1 is on their way. But barring something unforeseen, they won’t get their wish.
Assuming the Devils do opt for a combination of these three, however, I started digging through the last four seasons of NHL goaltending data. Now, I’m not nearly as concerned about the uncertainty as I initially thought I’d be. Here’s what it boiled down to:
Goaltending is an extremely volatile position. Mehta & Co. are betting that volatility works in their favor.
That might sound like a risky way to approach one of the most important positions on a hockey team. But I think it’s actually one of the more calculated decisions Mehta has made.
NHL Goaltending Is Harder to Predict Than It Seems
The easiest way to understand why Mehta’s approach makes sense is to look at what actually happens to NHL goalies from year to year.
For anyone unfamiliar with MoneyPuck’s GSAx, or Goals Saved Above Expected, it essentially measures how many goals a goalie prevented compared to what would be expected based on the quality of shots he faced. A positive number means he saved more goals than expected, while a negative number means he allowed more than expected.
GSAx/60 then shows how many goals above or below expected a goalie stopped per 60 minutes of ice time. It isn’t perfect, but it provides much more context than save percentage alone.
Looking at the last four seasons of NHL data for all goalies who played 25+ games, the year-to-year volatility is pretty remarkable.
For the purposes of this article, I’d roughly view GSAx/60 like this:
| GSAx/60 | Description |
|---|---|
| +0.30 or higher | Excellent |
| +0.15 to +0.29 | Very good |
| +0.05 to +0.14 | Good / above average |
| -0.04 to +0.04 | Around average |
| -0.05 to -0.14 | Below average |
| -0.15 to -0.29 | Bad |
| -0.30 or lower | Terrible |
And this isn’t just a matter of finding a few weird outliers. Across the three year-to-year transitions in the four-year sample, 20 goalies went from above-average GSAx/60 to below-average the following season, while 24 did the opposite.
In other words, there were 44 instances of a goalie crossing from one side of league average to the other in the following season. That’s more than half of the 83 goalies in the sample.
Here are just some examples that stood out:
Jordan Binnington went from -0.16 GSAx/60 in 2022-23 to +0.30 the following season, +0.10 in 2024-25 and -0.58 last year. That’s below average to excellent, back to above average and then all the way to terrible.
His GSAx tells an even crazier story. He went from roughly -9 to +16.5 to +5 to -22 over those four seasons. That’s a nearly 40-goal swing between his 2023-24 and 2025-26 seasons.
Karel Vejmelka had an equally strange path. His GSAx/60 went +0.23 → -0.20 → +0.25 → +0.11 from 2022-23 through 2025-26. He was good, bad, excellent and then good again.
Jeremy Swayman might be the clearest example of all. His GSAx/60 went from +0.72 in 2022-23 to +0.43 in 2023-24, before falling all the way to -0.16 in 2024-25 and then rebounding to +0.53 last season. In the span of three seasons, he went from excellent to bad to excellent again.
Even the great Andrei Vasilevskiy was impacted. After posting a huge positive GSAx total in 2022-23, he went all the way below zero in 2023-24 before bouncing back to another enormous positive season. And those are only a few examples. There were plenty more goalies who made similar swings over the sample.
There are obviously reasons behind the large fluctuations in individual seasons. Injuries, workload and plenty of other factors matter. But predicting exactly how a goalie is going to perform over the next 60 games is clearly more difficult than simply looking at his previous season and assuming he’ll do something similar.
There also isn’t some obvious relationship with age. John Gibson went from -0.23 (bad) and -0.22 (bad) in consecutive seasons to +0.59 (excellent) and +0.22 (very good).
Goalies just fluctuate. A lot. And that’s really the point.
For the Devils, utilizing additional cap space/assets to try to eliminate what is essentially still a wild card doesn’t necessarily make sense. No matter the age or track record, even the league’s best goalies can go from excellent to borderline unplayable in the span of a single season.
Situation is Already (Likely) an Upgrade Over Markstrom
Markstrom finished 2025-26 with a -11.9 GSAx and -0.281 GSAx/60, a dramatic drop from the 13.7 GSAx and +0.29 GSAx/60 he produced with Calgary in 2023-24.
The question isn’t whether Allen, Rittich or Daws are individually better bets than Markstrom — though they probably are. Rather, it’s whether the Devils would rather commit $6 million to a goalie who just finished 11.9 goals below expected, or instead use those resources to improve the roster while giving themselves multiple options in net.
Considering how last season went, I don’t think that’s a particularly difficult decision. And the league-wide data makes the decision even more understandable.
If a great goalie can quickly go from +16.50 GSAx to -22.39 — a ~38-goal swing, there is no such thing as a completely safe investment in the position.
The Devils’ Trio Has Seen This Volatility Firsthand
The Devils’ own goalies provide plenty of evidence. But the good news with a trio is that only one has to be above average and another serviceable to create a good situation.
Rittich finished 2023-24 with a .921 SV% and 13.8 GSAx, only to fall to an .886 SV% and -11.4 GSAx the following season. Then he bounced back last year, posting an .894 save percentage and +4.2 GSAx.
After Allen finished 2022-23 and 2023-24 with negative GSAx totals, he exploded to 18.4 GSAx and a .906 SV% in 2024-25 before remaining above average last season at 6.3 GSAx and .903%.
And then there’s Daws, who has never had the opportunity to establish himself over a full season — but has been tremendous in his short NHL stints.
While there’s no guarantee that any of them becomes a legitimate No. 1, the Devils don’t need one of them to suddenly become a star. Even if the trio cumulatively decreased by 20% from their 2025-26 production, they’d still produce +8.72 GSAx. For reference, with Markstrom’s struggles, the Devils finished with a total of -5.2 GSAx from those three goalies last season.
So even after accounting for a 20% decrease, that would represent a roughly 14-goal improvement over last season.
To put it all into perspective, the Devils scored a historically bad ~48 goals below expected last season because of their finishing. (via Natural Stat Trick) That’s nearly 10 times the impact of that -5.2 GSAx in net.
With a more complete lineup now, we can expect at least some positive regression from their shooting. If they can get that while replacing Markstrom’s negative impact in net with a trio that has already shown the ability to provide above-average goaltending, the entire equation starts to look very different.
And even with their finishing/goaltending issues last season, the Devils were 26-14-2 (.643 PTS%) with a fully healthy Jack Hughes in the lineup. If he’s able to stay healthy, that also dramatically changes things.
Allen Speaks on Three-Goalie System
These guys aren’t computers, either. They know when they’re feeling good. Coaches know when a goalie is fighting the puck. Players know when someone has earned another start.
The Devils can adjust accordingly as the season goes. If Allen is playing well, he can get more work. If Rittich is playing well, the same applies. If Daws takes a step forward, the Devils can increase his responsibility.
And Allen himself understands that.
In an interview with Daily Faceoff Live, Allen admitted that he didn’t particularly like playing in a three-goalie system with Montreal because it could make him feel “on the outs.” But he also explained that clear communication between the goalie coach, head coach and goalies can make the setup work. He called himself a realist and acknowledged that teams need multiple capable goalies in today’s NHL.
“There’s just so much variability,” Allen said. “Every game is so important.”
The Devils aren’t necessarily asking three goalies to split the season evenly because they think that’s some revolutionary strategy. Instead, they’re giving themselves options in a position where, as Allen put it, there is so much variability.
Mehta Standing True to His Philosophy
Since taking the job, Mehta has been adamant that there is no one specific way to build a championship team.
At his introductory press conference, he described the GM’s job as taking in information, assessing risk and making decisions despite the inherent uncertainty.
“Data does undoubtedly give a huge advantage in hockey in terms of projecting future performance of players, but you have to understand all those same things, the character, the locker room, the culture, the intangibles,” he said. “All that stuff matters.”
That’s essentially what we’re seeing with the goaltending situation. There isn’t one formula that says a Stanley Cup contender needs a $7 million starter — or a 60-game workhorse.
And there certainly isn’t one that says the goalie with the biggest contract will provide the best performance. The goalie market is extremely thin. That’s part of why established starters cost so much in the first place. When a team has the opportunity to acquire someone it trusts, there’s enormous pressure to pay for the security of knowing who will be in the crease.
But the contract doesn’t automatically eliminate the risk. Markstrom is direct proof of that. So are Binnington, Swayman, Vejmelka, Gibson and plenty of others.
Where Mehta’s Background Matters
It’s no coincidence that Mehta seems comfortable with this. Before getting into hockey management, Mehta worked as an options trader and played poker professionally. He later spent three seasons as an assistant GM with the Florida Panthers, helping them win back-to-back Stanley Cups. His background is essentially built around making decisions when the outcome isn’t guaranteed.
You don’t need to know exactly what is going to happen, but you do need to understand the odds, manage the downside and put yourself in position to benefit when things break your way.
That’s what this all feels like:
He has three goalies who have great upside, information that will accumulate throughout the season, and a coaching staff capable of adjusting workloads. Furthermore, in future seasons, he’ll have young goalies like Jakub Malek and Mikhail Yegorov who could eventually create inexpensive options for the organization. That all bodes well for not just this season, but Mehta’s goal of building a sustainable, long-term contender.
On top of it all, he still has approximately $3.85 million in cap space available. On the off chance the situation becomes untenable, he can make a move.
Why the Gamble Is Worth Taking
The Devils’ new goaltending situation absolutely carries risk. All three could struggle. But the opposite could also be true.
What if Allen gives New Jersey another season resembling his +18.4 GSAx campaign?
Rittich could continue the positive trend he established last season. Daws could build off his limited NHL success and show why the organization remains so high on him.
Even if one of those three things happens, they’re probably in a pretty good spot — all things considered. That’s the bet. The four-year sample suggests it’s not nearly as crazy as it sounds.
Mehta warned Devils fans about this possibility from the beginning.
“I’m up here in this chair now, and there’s no doubt in my mind that there’s gonna be times where fans disagree with me, and there’s gonna be times where they might question my decisions, but the one thing they’ll never be able to question is my sincerity, my passion and my loyalty to this franchise.”
Considering how often NHL goaltenders swing from great to bad, bad to great and everything in between, I think Mehta has made a pretty reasonable decision to let the season tell him who deserves the net...all while allocating resources to areas that needed it more.
Mantha is a perfect example. His +13.43 goals above expected last season would have led the Devils; his positive impact would’ve been nearly three times as large as the negative impact from their -5.2 GSAx goaltending.
There are certainly reasons to be skeptical of any plan that goes against the norm. But after digging through the numbers, Mehta’s calculated gamble looks far more reasonable than it appears on the surface.