Sorokin can't carry this again

On April 4 the Islanders sat third in the Metropolitan Division. They won four of their last 14 games, fired Patrick Roy with four to play, and finished sixth. Along the way the club produced the unanimous Calder Trophy winner, a 19-year-old defenseman with 23 goals, and the Vezina runner-up in Sorokin.

So the bettor's question for 2026-27 is the one the collapse raises. Were the Islanders a playoff team that ran out of gas, or a weaker team that Sorokin carried for five months?

Our answer is the second. New York was out-chanced most nights and finished its chances worse than almost anyone. Its goalie covered all of it. Goaltending at that level does not repeat reliably, and the summer did little to make the skaters better. Anders Lee left, Matias Maccelli arrived, and the rest of the lineup got a year older.

That puts us close to the gloomier corner of the crowd. The Athletic's projection has the Islanders last in the East with about 82 points, and only two of 14 NHL.com staffers sneak them into a wild-card spot. The sportsbooks are kinder, and VSiN's preview likes the over on 88.5 on the strength of last season's 91 points. We think those 91 points flatter the team. Everyone agrees Schaefer and Sorokin are the reasons to watch. We agree too. Two great players have not been enough before.

1What last season really was

The record was middling. The play underneath it was worse, and the goaltending was the best in the league.

This is a familiar Islanders shape. Their shooters have finished below expected goals in three of the last four seasons, and the two best years in that stretch came when Sorokin posted league-best numbers. Expected goals measure the quality of each chance from its location and type, so finishing below them means the shots were there and the goals were not. The weak finishing looks like a real trait of this forward group. Elite goaltending is real too, but its size swings a lot from year to year.

The fade followed the net. The team took 48 points in the first 41 games and 43 in the last 41, and its goaltending edge shrank by almost half between the halves. Sorokin was spectacular from November through January and close to average from February on, while David Rittich slipped below expected. The chances against never went away. In the spring the saves stopped covering them.

2How the model builds New York's ranking

Our preseason rating starts from how well a team actually played last season, with luck stripped out, and then adds only what has historically carried over from one season to the next.

The starting point is below the standings. Take out the part of last season's goaltending and finishing that tends not to repeat, and the Islanders played like a bottom-ten team, not a 91-point one. This is the core of our view.

The roster pulls it down further. Losing the captain's net-front scoring with only a middle-six playmaker coming in is a net loss. The forward group also leans on 35-year-olds (Brayden Schenn, Kyle Palmieri, Ondrej Palat, Casey Cizikas) and on a 19-year-old rookie in Victor Eklund. Barzal starting the season hurt adds to the drag.

Finishing and goaltending cancel out. Over three seasons, the Islanders' poor shooting and strong goaltending roughly offset each other, so neither earns a bonus or a penalty.

Age is a small minus. The lineup is a bit older than the league average, and older teams tend to slip. Schaefer, Ritchie and Eklund keep it from being worse.

Summer moves: in and out

Net: a slightly weaker and older lineup, with a weaker backup goalie. The only notable addition is Maccelli. Defenseman Logan Stanley signed a one-year deal on the eve of the season and adds size as the eighth defenseman.

Two veterans arrived during last season and now count for a full year: Palat from New Jersey in January and Schenn from St. Louis at the deadline. Schenn centers the second line. Both are 35, so this is a stopgap more than an upgrade.

The coach is a small plus. DeBoer took over for the last four games of last season. His teams have been slightly better than their rosters over a long career, but new coaches get only modest credit in our model, because most of a coach's measured effect does not travel with him.

The crease barely moves it. Sorokin is back, which the market and our model both know. The backup downgrade costs a little.

The result is a D+ team, 25th in the league, projected for about 83 points. That puts the Islanders in the East's bottom group, below the playoff line and a few points under the sportsbooks. Books usually beat preseason models on season totals, so we hold that view loosely. The gap is small enough that a healthy Barzal and another Sorokin year could close it.

Our projected opening-night lineup. Each player shows his 2026-27 grade (A to E, with a 0-100 score: his percentile among NHL regulars at his position), his role and last season's numbers. Goalies are graded the same way against the league's 64 starters and backups.

Reading the lineup

Two centers carry the forwards. Barzal grades A and Bo Horvat B+. Barzal led the team with 72 points as the playmaker, and Horvat scored 31 goals in 68 games. With Barzal out, Anthony Duclair moves up beside Horvat and Simon Holmstrom, and the top line loses most of its creativity. Barzal's knee is the single biggest variable in the October price.

The weak spot is scoring depth. Our lineup score ranks the Islanders' top six among the thinnest in the league. Schenn and Palmieri are 35, Calum Ritchie and Eklund are still learning, and Emil Heineman scores mostly from the front of the net. The common view is that the offense depends on Schaefer and a better power play. Our numbers agree: there is not much else to lean on.

The bottom six can defend. Jean-Gabriel Pageau and Cizikas are experienced checkers, and Pageau still kills penalties. This group will not lose many games on its own. It will not win many either.

The blue line is the strength. Schaefer grades A and plays nearly 25 minutes a night as a teenager. He was the unanimous Calder winner, and with 23 goals and 59 points as a rookie he is a natural early Norris talking point. Pelech grades B and does the quiet work. Behind them the grades drop fast: Pulock, Tony DeAngelo and Scott Mayfield all grade C- or lower, and Alexander Romanov played only 15 games last season.

4Who matters most

Fans will look at Barzal and Sorokin first, and they are right to. Our absence research adds the defensemen to that list. A defenseman who plays big minutes matters more to the nightly price than his point total suggests, because his minutes go to someone worse. On this roster the drop from Schaefer and Pelech to the third pair is steep, so a missing top-four defenseman costs the Islanders more than it would cost a deeper team.

The market reacts quickly to star forwards and goalies. It moves much less when a defenseman sits. On game day, check the blue line as well as the crease.

5Goaltending: the best in the league, and the thinnest behind him

Sorokin finished second in Vezina voting and led the league in shutouts. Rittich was steady behind him. Together they saved more goals above expected than any tandem in the league, and the monthly table shows where the season was won: Sorokin's November through January. GSAx, goals saved above expected, compares the goals a goalie allowed with the quality of the shots he faced.

Our projection agrees with the crowd on the starter. Sorokin grades A+ and ranks among the very best goalies in the league. Daily Faceoff ranks the Islanders' tandem seventh, which fits. Our caution is about the size of last season. Even the best goalies rarely repeat a year that far above expected, and our projection pulls him back toward a very good normal. That regression is the main reason last season's points do not carry over.

The backup is the real question. Semyon Varlamov is 38 and has not played an NHL game since November 2024, after surgery on both knees. He looked sharp in the preseason and won the job over Vitek Vanecek, who went to the AHL. Our projection grades Varlamov at the bottom of the pool, mostly on age and two lost seasons. With 13 back-to-back sets on the schedule, he will start games. Those are the nights when the Islanders look like a much weaker team. Confirm the starter before you bet.

6New York's system: how they score and defend

Peter DeBoer's teams have a clear signature: they keep the puck, take more shots than their opponents and give up fewer. The Islanders under Roy did not look like that. They were even on shot volume but gave up the most rebound attempts in the league and the second-most attempts from the slot. Sorokin erased most of it. DeBoer's job is to take that work off his goalie.

Offense: volume without danger. The Islanders got plenty of rebounds but few shots from the slot, and a smaller share of their goals came from the slot than on almost any other team. That is how a team ends up below its expected goals year after year. DeBoer's teams create more from possession, which should help at the margins.

Defense: too much in front of the net. Rebounds and slot chances against are the costliest shots in hockey, and the Islanders gave up more of both than nearly anyone. They allowed few goals from them only because of Sorokin. If DeBoer cleans up the slot, the team gets better even if Sorokin comes back to earth.

What it means for scoring. Islanders games averaged 5.67 total goals last season: 2.79 for and 2.88 against per game, against a league average of 3.08 per team (6.16 per game). Our team scoring ratings carry that into 2026-27: a below-average offense and a middle-of-the-pack defense, which makes the Islanders one of the lower-event teams in the league. Tight, low-scoring games are the default.

7Special teams

The power play was one of the worst in the league on goals and on the chances it created, so bad luck does not explain it. The penalty kill ranked among the best on goals allowed while allowing middling chances, which is Ilya Sorokin's doing more than the system's.

How the power play works. The most interesting piece is the youngest. Schaefer put up 15 power-play points and eight goals from the point as a rookie, rare production for a teenage defenseman, and the unit now runs through him. Horvat works the middle and the net front, Palmieri shoots from the flank, and Schenn, a deadline pickup, adds a veteran body in front. Maccelli arrives from Toronto to add the half-wall passing the group lacked.

A new coach for it. The club hired Rocky Thompson to run the power play, and he inherits more skill than last season's group had. Losing Anders Lee's net-front work hurts, and Barzal's knee leaves the second unit without its best playmaker until he returns. Structure and personnel both have room to improve.

How the penalty kill works. Pulock and Pelech are one of the league's steadiest shutdown pairs, and they carry the kill on the back end. Pageau takes the key faceoffs and Holmstrom adds speed up front. The group allowed chances at a middling rate and Sorokin erased many of them.

What to watch. Two things point in opposite directions. The power play has nowhere to go but up, especially once Barzal returns. The penalty kill's goals against were better than its chances, and that gap tends to close even with an elite goalie. Expect both units to finish closer to average.

8What would change our view

• Barzal's kneeA quick return restores the top line. A lingering problem makes the forward group one of the thinnest in the league.
• The slotIf rebounds and slot chances against drop under DeBoer by November, the team is better than its rating, whatever Sorokin does.
• VarlamovIf he holds up and plays like an average backup, the weak nights get less weak. If his knees give out, Vanecek is the fallback.
• The power playA new coach for the unit and more skill on the point. A move to the middle of the league would add real goals.
• Schaefer in year twoA step forward on defense would lift the whole blue line. Second-year dips happen too.
• Early-season formThe first 20 to 30 games say less than they seem. Hot and cold starts mostly reflect shooting and goaltending swings, and this team lives on goaltending.