Training Camp 2026: A Realistic Lineup for All 32 Teams, and the Questions Camp Has to Answer

Camps are open. I started with the Whiteboard's lineup model: its suggested best lineup for every team, built from two analytics. Drive measures five-on-five chance flow, how the puck tilts with a player on the ice. Engine measures individual chance creation and finishing. Chemistry blends playing-style tag fit, five-on-five complementarity, and whether each player fits the line's role, on a 0 to 100 scale.

Then I did the human part. I adjusted for injuries, put players back in their real positions, and lined things up the way camp reporting suggests coaches will ice them. Then I ran every adjusted lineup back through the Whiteboard's chemistry engine, the same math the tool uses when you drag a winger onto a new line and watch the number move. Every score below is the model's own math on a realistic lineup.

For each team: the lineup, what the numbers say about the roster, and the questions camp has to answer.

Anaheim Ducks

The lineup:

• Forwards
• L1: Cutter Gauthier - Leo Carlsson - Beckett Sennecke (79)
• L2: Frank Vatrano - Ryan Poehling - A.J. Greer (61)
• L3: Alex Killorn - Mikael Granlund - Jeff Malott (62)
• L4: Nikita Nesterenko - Tim Washe - Sam Colangelo (35)
• Defense
• P1: Jackson LaCombe - Tristan Luneau (54)
• P2: Pavel Mintyukov - Nick Jensen (60)
• P3: Ian Moore - Drew Helleson (39)

Goalies: Lukas Dostal (Strong, 66), Ville Husso (Average, 46)

Troy Terry is out until at least November after surgery on June 9 to repair a hip impingement and labral tear, and he could miss into December. He is a top winger and a leader on this roster, and losing him for the first two months costs a team that made its first playoff run since 2018 any veteran steadiness up front. Terry played through the injury for half of last season and still finished with 57 points in 61 games.

The other pillar of this camp is how much of the roster turned over. Mason McTavish was traded to St. Louis, Radko Gudas went to Florida in the deal that brought in A.J. Greer, Jacob Trouba left for San Jose, John Carlson was dealt to Carolina, and Olen Zellweger went to Buffalo. Greer, Jeff Malott, and Nick Jensen are the incoming pieces. The team that broke the playoff drought is partly gone before the new 84-game season starts.

What is left at the top is exciting. Cutter Gauthier led the team with 41 goals and 69 points last season, Leo Carlsson grades as a strong driver, and Beckett Sennecke grades one too. Their line's chemistry score is 79 out of 100, which measures how well three players' styles fit together, and it is the best forward unit on the roster by a distance. Behind it, the drop is steep: the second line grades 61, and the fourth line grades 35, which is replacement level. Greer is worth a look on that second line, because his drive rating sits at 88, which measures how much five-on-five chance flow tilts when a player is on the ice, the best mark of anyone on the bottom three lines.

On defense, one pair the model trusts: Mintyukov - Jensen grade 60, driven by Jensen's veteran steadiness. The top pair of LaCombe and Luneau grades 54 and is still getting to know each other. After that, nothing.

Then the crease. Based on the model, Lukas Dostal grades 66, which puts him in the Strong tier. He saved 0.34 goals above expected per 60 minutes last season, which measures how many goals he prevents compared to an average goalie facing the same shots, and stopped 93 percent of rush chances. Ville Husso grades 46, average, and gives a few goals back by the same measure. The depth chart is clear: Dostal is the starter.

Camp questions: when Terry returns, does he walk back onto a top line, and whose spot does he take? Can LaCombe and Luneau become a real top pair by November? And who scores when the kids sit, because the bottom six grades like one.

Boston Bruins

The lineup:

• Forwards
• L1: JJ Peterka - Pavel Zacha - David Pastrnak (77)
• L2: Morgan Geekie - Fraser Minten - Casey Mittelstadt (51)
• L3: Michael Eyssimont - Elias Lindholm - Mark Kastelic (43)
• L4: Tanner Jeannot - Sean Kuraly - Alex Steeves (56)
• Defense
• P1: Jonathan Aspirot - Charlie McAvoy (49)
• P2: Hampus Lindholm - Will Borgen (56)
• P3: Nikita Zadorov - Connor Clifton (50)

Goalies: Jeremy Swayman (Strong, 78)

The first wild card in the East took the Atlantic Division champions to six games in April, and the model thinks that was not a fluke. The top line grades 77 chemistry, which measures how well three players' styles fit together, and it is easy to see why. Zacha is back at center after a career year of 30 goals and 65 points in 78 games. Pastrnak carries elite playmaker and dual-threat tags. Peterka, acquired from Utah on draft night for two first-round picks, brings a 25-goal season and an elite trigger-man tag. That is a top line, not a prayer.

Here is the depth test. Average quality across the four lines runs 63.7, 50.8, 32.6, 38.3, which means the fourth line out-grades the third: Jeannot, Kuraly and Steeves all carry drive ratings above 60, and drive rating measures how much five-on-five offense flows through a player. The pairs are even tighter: 49, 56, 50, three usable pairs within seven points of each other. Hampus Lindholm grades 79.8 quality with a drive rating of 84, the best defenseman on the roster by the model's numbers, and McAvoy grades 71. That is not a top-heavy roster. That is a roster with a real top line, three balanced pairs, and bottom-six forwards who create more than they cost.

Then there is the crease. Based on the model, Jeremy Swayman grades 78, Strong. He stopped 0.47 goals above expected per 60 minutes last season, meaning he saved roughly half a goal a game that an average goalie would have allowed, and turned away 93.0 percent of rush chances. A Strong goalie behind three balanced pairs is how wild-card teams become second-round teams.

The honest risks: the third line grades 32.6 and Elias Lindholm sits at a 23.6 quality rating, the model's polite way of saying the middle of the lineup has a soft spot. Fraser Minten, at 47.6 quality, is being asked to center the second line. And Zacha is in the final year of his contract, with extension talks still going as camp opens. The rebound case: a real top line, depth that grades better than its reputation, and a real goalie. The bet is that the third line does not sink it.

Camp questions: can Minten hold the second-line center job, or does Mittelstadt slide over? Does Zacha sign before the season opens? And can the fourth line keep out-grading the third, because 32.6 on L3 is the number that could sink the rebound?

Buffalo Sabres

The lineup:

• Forwards
• L1: Zach Benson - Josh Norris - Tage Thompson (90)
• L2: Josh Doan - Jiri Kulich - Jack Quinn (86)
• L3: Jason Zucker - Ryan McLeod - Peyton Krebs (64)
• L4: Conor Sheary - Noah Ostlund - Justin Danforth (27)
• Defense
• P1: Rasmus Dahlin - Mattias Samuelsson (60)
• P2: Owen Power - Conor Timmins (29)
• P3: Olen Zellweger - Louis Crevier (61)

Goalies: Alex Lyon (Strong, 81), Ukko-Pekka Luukkonen (Strong, 60)

Buffalo owns two of the best-graded lines in the league, and the model is not subtle about it: the top line scores 90, the second 86. Chemistry grades how well three players' styles fit together, and 90 is about as high as the scale goes. The top unit is Benson (quality 75.7, drive rating 90) with Norris (72.2, drive rating 90) and Thompson, who scored 40 goals and 81 points in 81 games last season. Drive rating measures how much five-on-five offense flows through a player, and two 90s plus a 40-goal scorer is a lot of flow. The second line is Doan, an elite driver with a drive rating of 98, flanking Kulich and Quinn. Two lines, five drivers, one sniper. Then the floor falls out: 64, then 27. The fourth line grades like a preseason roster.

The blue line tells the same story in miniature. Dahlin grades 95.0 quality, an elite driver with a drive rating of 94, after a 74-point season in 77 games. Samuelsson holds the top pair together at 60. Then the model gives up: Power with Timmins grades 29, while the third pair, Zellweger with Crevier, grades 61, better than the second pair. A 29 next to Dahlin's 60 is the kind of problem the playoffs find, and last spring they did: Montreal beat Buffalo in seven in the second round.

Then there is the crease, and here the model has a surprise. Based on the model, Alex Lyon grades 81, Strong, after stopping 0.46 goals above expected per 60 minutes and 96.5 percent of rush chances in 36 games. Ukko-Pekka Luukkonen grades 60, also Strong, with 0.22 above expected per 60 in 35 games. Two goalies grading Strong is a genuine tandem, and on a roster whose second pair grades 29 and whose fourth line grades 27, the crease is not a luxury. It is load-bearing. If one of them holds, Buffalo can survive the minutes the bottom of the roster cannot.

Camp questions: does Kulich hold the second-line center job? Does anyone fix the second pair, because 29 next to Dahlin will not survive April? And which goalie gets opening night, Lyon or Luukkonen?

Calgary Flames

The lineup:

• Forwards
• L1: Matvei Gridin - Morgan Frost - Matt Coronato (55)
• L2: Maxim Tsyplakov - Mikael Backlund - Joel Farabee (52)
• L3: Connor Zary - Ryan Strome - Samuel Honzek (59)
• L4: Aydar Suniev - Martin Pospisil - Adam Klapka (46)
• Defense
• P1: Jake Middleton - Simon Nemec (59)
• P2: Kevin Bahl - Zach Whitecloud (62)
• P3: Joel Hanley - Zayne Parekh (38)

Goalies: Dustin Wolf (Strong, 65), Devin Cooley (Elite, 84)

Calgary is fully rebuilding, and the roster shows it. Nazem Kadri was traded back to Colorado at the deadline, Blake Coleman and MacKenzie Weegar went out too, and the franchise is stockpiling first-round picks instead of chasing veterans. The numbers are blunt about what remains: no line grades above 59, and no unit on this roster is being asked to win anything this year.

Jonathan Huberdeau will not be ready for the season opener either. He had hip resurfacing surgery in March, suffered a setback in his rehab this month, and is starting camp skating on his own. He managed 25 points in 50 games before being shut down, and whatever he has left is a next-season question.

The best underlying skater on the roster is Maxim Tsyplakov, a strong driver with a drive rating of 94, which measures how much five-on-five chance flow tilts when a player is on the ice. He slots onto the second line, where he belongs. Around him it is all kids: Gridin, Coronato, Zary, Honzek. Nobody here has ever carried a top line, and the model's math says nobody here is about to start. The third line grades 59, the best mark on the roster, which tells you everything about the state of the offense.

On defense, the new faces carry the load: Jake Middleton and Simon Nemec, both acquired this summer, pair up at 59, while Kevin Bahl and Zach Whitecloud grade 62 as the best pair on the team. Frost and Whitecloud are already drawing trade chatter around the league, which is what happens in a rebuild. The third pair is where the future gets its audition: Zayne Parekh, the ninth overall pick in 2024, skates with Joel Hanley in the model's lineup. The model is blunt about him: 6th percentile and a Passenger grade after a rookie year that never got going, and the pair grades 38. He is in the NHL group at camp, and the rebuild gives him every chance to take the job. Whether he is ready for a real role or just a roster spot is the biggest question on the blue line.

The crease is the one bright spot, and it is genuinely bright. Based on the model, Devin Cooley grades 84, which puts him in the Elite tier, saving 0.71 goals above expected per 60 minutes. Dustin Wolf grades 65, Strong, and stopped 96.4 percent of the rush chances he faced. The backup grades higher than the starter, which is a good problem in a season where wins are not the point.

Camp questions: which of the kids seizes a top-six job and keeps it, and is Parekh ready for a real role on the blue line? Does Huberdeau's setback change the timeline, or does the team simply wait? And is there a veteran addition coming, or is this openly a development year?

Carolina Hurricanes

The lineup:

• Forwards
• L1: Nikolaj Ehlers - Sebastian Aho - Andrei Svechnikov (82)
• L2: Jackson Blake - Logan Stankoven - Taylor Hall (82)
• L3: William Carrier - Jordan Staal - Jordan Martinook (70)
• L4: Eric Robinson - Mark Jankowski - Jesperi Kotkaniemi (63)
• Defense
• P1: Jaccob Slavin - Jalen Chatfield (53)
• P2: K'Andre Miller - Sean Walker (67)
• P3: Shayne Gostisbehere - Mike Reilly (49)

Goalies: Brandon Bussi (Strong, 62.2), Pyotr Kochetkov (Average, 52.5)

Alexander Nikishin will not be at camp. The 24-year-old defenseman is an unsigned restricted free agent, and general manager Eric Tulsky said on September 16 that he will not report without a deal. Per reporting, Nikishin asked for a trade after the Cup run and has not backed off. He put up 33 points as a rookie and finished seventh in Calder voting, then saw his ice time shrink to about 15 minutes a night in the playoffs behind Jalen Chatfield and Sean Walker. He is not in the lineup below because he is not under contract, and his absence is the biggest storyline in Raleigh. Whether this ends in a signature or a trade shapes the blue line for years.

Seth Jarvis is out too, after shoulder surgery in June, with the team saying four to six months, so he misses at least the first month of the 84-game season. Carolina's answer is the deepest forward group in the division. Both top lines grade 82 on the model's chemistry scale, which measures how well three players' styles fit together, the best top-six marks in the Metro. In the model's lineup, Nikolaj Ehlers, the summer's top free agent, flanks Sebastian Aho, an elite driver at 83.7 quality, with Andrei Svechnikov on the first line. Logan Stankoven, who led the team with 11 goals during the Cup run, centers the second between Jackson Blake and Taylor Hall. All four lines grade 63 or better, something only Pittsburgh matches in the division.

The worry is the blue line. Jaccob Slavin is still elite, but his pair with Chatfield grades only 53. The best unit is the second: K'Andre Miller, acquired from the Rangers, with Sean Walker at 67. Walker's drive rating of 74 measures how much five-on-five offense flows through a defenseman, and it is doing a lot of the lifting there.

Then the crease, and it is the post-Andersen era. Frederik Andersen signed with Edmonton on July 1 after going 13-2 with a 1.89 goals-against average in the playoffs. Based on the model, Brandon Bussi grades 62.2, which is Strong, and Pyotr Kochetkov grades 52.5, Average. Bussi stopped 0.25 goals above expected per 60 minutes last season, which is solid work in 39 games, and he finished the Final after Andersen's knee injury. But a full season as the starter is a different job, and behind him sits a 52.5. The honest read: the model thinks this crease can work, but it is the thinnest it has been in years.

Camp questions: does Nikishin sign or get traded, and who plays his minutes if he sits? When Jarvis returns, whose spot does he take on a team with four good lines? Is Bussi a real No. 1 over 84 games?

Chicago Blackhawks

The lineup:

• Forwards
• L1: Tyler Bertuzzi - Anton Frondell - Patrick Kane (53)
• L2: Jordan Greenway - Frank Nazar - Nick Lardis (43)
• L3: Oliver Moore - Ryan Donato - Cole Smith (50)
• L4: Ryan Greene - Teuvo Teravainen - Andrew Mangiapane (45)
• Defense
• P1: Bowen Byram - Alex Vlasic (68)
• P2: Ian Cole - Wyatt Kaiser (49)
• P3: Ethan Del Mastro - Kevin Korchinski (23)

Goalies: Spencer Knight (Elite, 82), Arvid Soderblom (Below avg, 23)

Connor Bedard had left shoulder surgery on July 8 and the team says four months, which puts him back in early November and costs him the first month of the 84-game season. He led the team with 75 points in 69 games last year, so this is not a small absence. It is also why the first-line center in the model's camp lineup is Anton Frondell, the 19-year-old 2025 No. 3 overall pick, and why Patrick Kane is on the first line at all.

Patrick Kane starts on the first line, and the model has a clear read on him: 31 points at five on five in 67 games last season, strong Playmaker and Secondary scorer tags, and a quality rating at the 47th percentile, which the model calls Neutral. In plain terms, that means the model sees a player who still moves the puck and finishes at a second-line rate but no longer drives play the way he did in his prime. The 47th percentile is not an insult and not a compliment. It is the middle of the league, and at his age that is roughly what the tape says.

His line with Bertuzzi and Frondell grades 53 out of 100 on the model's chemistry scale, which measures how well three players' styles fit together. It is the best forward unit on the roster, mostly because Bertuzzi's net-front game and Kane's passing simplify things for a rookie center. The best unit overall is the top defense pair. Bowen Byram and Alex Vlasic grade 68, and the reason is Byram: his drive rating sits at 96, which measures how much five-on-five offense flows through a defenseman, and Vlasic is the steady partner who lets him do it.

Then there is the crease. Based on the model, Spencer Knight grades 82, right at the Elite line. He stopped 0.42 goals above expected per 60 minutes last season and turned away 97.8 percent of the roughly 45 rush chances he faced, which is the kind of number that separates good goalies from expensive ones. Arvid Soderblom grades 23, below average. The depth chart is simple: Knight is the starter, and it is not a competition.

One data note: Roman Kantserov is penciled for a top-nine look at camp but has no NHL data in the model yet, so he does not appear in the numbers above.

Camp questions: does Frondell look like a future first-line center or a teenager keeping a seat warm? When Bedard returns, does Kane slide onto his wing and does this team suddenly have two real lines? And is Knight enough to steal the games a rebuilding roster will keep losing?

Colorado Avalanche

The lineup:

• Forwards
• L1: Artturi Lehkonen - Nathan MacKinnon - Martin Necas (88)
• L2: Gabriel Landeskog - Nazem Kadri - Nicolas Roy (56)
• L3: Logan O'Connor - Brock Nelson - Zachary L'Heureux (63)
• L4: Jaden Schwartz - Fedor Svechkov - Parker Kelly (60)
• Defense
• P1: Devon Toews - Cale Makar (77)
• P2: Brett Kulak - Sam Malinski (48)
• P3: Brent Burns - Josh Manson (69)

Goalies: Scott Wedgewood (Strong, 78), Mackenzie Blackwood (Strong, 66)

Start with how good this team is, because the numbers demand it. Colorado won the Presidents' Trophy at 55-16-11 and led the league in goals per game (3.63), fewest goals against per game (2.40), and penalty killing (84.6 percent). The one category where they were ordinary was the power play: 27th in the league at 17.1 percent. That is the lone blemish on a historically good regular season, and it is the entire story of this camp.

The top line is why they are the best team in hockey. Necas and MacKinnon played together, and it was as good as it sounds: the two combined for 227 points, 91 goals and 136 assists. Necas posted a career year in his first full Colorado season, 38 goals and 100 points, finishing sixth in the league in even-strength points, then signed an eight-year extension in October. MacKinnon scored 53 goals with 74 assists. In the model, MacKinnon's drive rating is 100, which measures how much five-on-five offense flows through a player, with elite Playmaker, Sniper, and Play driver tags. Lehkonen brings elite net-front and secondary-scorer tags. The line grades 88 out of 100 on the model's chemistry scale, which measures how well three players' styles fit together, at an average quality of 90.9. That is the class of the division.

Now the power play. Makar carries an elite power-play quarterback tag and the unit still finished 27th. That is not a personnel problem. It is a systems problem, or a luck problem, or some mix of the two, and camp will show whether anything about the setup changes.

Kadri came back to Colorado in the March 6 trade from Calgary and produced right away. He put up four goals and nine points in his first 12 games back, six of them on the power play, starting on the wing because MacKinnon and Brock Nelson were already down the middle. The model's read is more sober: Neutral quality at the 55th percentile, drive rating 20, meaning he no longer drives five-on-five offense. The model's lineup has him as the second-line center between Landeskog, a strong driver with a drive rating of 88, and Nicolas Roy, who grades as a passenger. It is an honest second line, not a great one.

The depth is real behind it. Nelson, an elite driver with strong sniper and play-driver tags, anchors the third line at 63 chemistry. On defense, the third pair out-grades the second: Burns and Manson, both elite drivers, grade 69 together, while Kulak, a passenger at the 34th percentile, drags Sam Malinski, an elite driver at the 93rd percentile coming off a career 40-point season, down to 48. That is a camp problem with a camp solution.

Then there is the crease. Based on the model, Scott Wedgewood grades 78, Strong. He stopped 0.58 goals above expected per 60 minutes last season, which is excellent, and turned away 95.9 percent of the rush chances he faced. Mackenzie Blackwood grades 66, also Strong, at 0.40 goals saved above expected per 60. There is no controversy here. There are just two good goalies.

Camp questions: can the power play climb out of the bottom five with an elite quarterback already in place? Is Kadri's line a real second unit, or does the five-on-five drop-off show? And does the third pair keep out-grading the second, or does someone fix the Kulak problem?

Columbus Blue Jackets

The lineup:

• Forwards
• L1: Valeri Nichushkin - Danton Heinen - Kirill Marchenko (74)
• L2: Dmitri Voronkov - Sean Monahan - Mathieu Olivier (71)
• L3: Cole Sillinger - Charlie Coyle - Conor Garland (42)
• L4: Miles Wood - Kent Johnson - Ryan Lomberg (23)
• Defense
• P1: Zach Werenski - Dante Fabbro (53)
• P2: Ivan Provorov - Damon Severson (57)
• P3: Denton Mateychuk - Jake Christiansen (44)

Goalies: Jet Greaves (Elite, 93.5), Elvis Merzlikins (Average, 43.6)

Adam Fantilli did not show up for the start of camp. The 22-year-old is an unsigned restricted free agent, and his agent has said he will not take part without a contract. General manager Don Waddell says he is confident a deal gets done, but the market reset hard this summer: Anaheim matched an $18 million average annual value offer sheet for Leo Carlsson, and San Jose gave Cutter Gauthier a six-year extension at $13.5 million per year. Fantilli put up 59 points last season, a career high, and Columbus opens the season October 1 at home against Buffalo. Every day without him is a day with Danton Heinen at first-line center.

Heinen is a useful player grading Neutral at 42.7 quality, but he is not a No. 1 center, and the lineup below is honest about that. The top line survives because of the wings: Valeri Nichushkin, the summer trade pickup, grades as an elite driver at 81.2 with a drive rating of 100, which measures how much five-on-five offense flows through a player, and Kirill Marchenko is a strong driver at 79.8. That line grades 74 on the model's chemistry scale, which measures how well three players' styles fit together. Sean Monahan's line grades 71 right behind it, and Monahan grades as a strong driver at 70.2. Then the cliff: the third and fourth lines grade 42 and 23, and Kent Johnson, grading as a passenger at 29.4, is on the fourth.

The step-up has to be collective. Nobody replaces Fantilli at center, but Nichushkin and Marchenko are good enough on the wings, and Monahan is steady enough in the middle, that the top six works without him. Zach Werenski, an elite driver at 97.0 with a drive rating of 95, remains the best player on the roster. The question is whether anything below the second line can score.

The crease, at least, is settled. Based on the model, Jet Greaves grades 93.5, Elite, the best goaltending mark in the division. He stopped 0.66 goals above expected per 60 minutes last season, which is the kind of number that steals games. Elvis Merzlikins, who grades 43.6 Average, had shoulder surgery in August and is out long term. The net is Greaves's, and the model thinks he is ready for it.

Camp questions: does the holdout end before October 1? If Fantilli signs, does Heinen slide to the third line, and what happens to Cole Sillinger? Can the bottom six score enough for the top six to matter?

Dallas Stars

The lineup:

• Forwards
• L1: Jason Robertson - Wyatt Johnston - Mikko Rantanen (74)
• L2: Joel Kiviranta - Roope Hintz - Jamie Benn (72)
• L3: Justin Hryckowian - Matt Duchene - Sam Steel (61)
• L4: Oskar Bäck - Tyler Seguin - Arttu Hyry (49)
• Defense
• P1: Miro Heiskanen - Esa Lindell (58)
• P2: Thomas Harley - Tyler Myers (37)
• P3: Nils Lundkvist - Kyle Capobianco (23)

Goalies: Jake Oettinger (Strong, 62), Casey DeSmith (Elite, 87)

Dallas has a reputation for great defense, and the forwards do nothing to contradict the contender talk: Robertson, with a drive rating of 94 and an elite play-driver tag, between Johnston and Rantanen, who carries an elite clutch tag, grades 74 out of 100 on the model's chemistry scale, which measures how well three players' styles fit together. The second line grades 72 behind Hintz. Scoring is not the problem in Dallas.

The blue line is the problem, and it is worth going pair by pair, because the reputation does not survive the numbers.

The top pair grades 58. Heiskanen is genuinely elite by the model: 92nd percentile quality, drive rating 99, an elite power-play quarterback tag and a strong suppressor tag. Lindell is a strong driver at the 79th percentile. Individually they are fine. As a pair, the fit is only middling, and on a contender the top pair cannot be middling.

The second pair grades 37. Harley is a strong driver by quality at the 71st percentile, but his tag sheet is thin: strong distributor, strong shot blocker, and nothing about driving play or defending. Myers is a passenger at the 37th percentile with a drive rating of 19, which measures how much five-on-five offense flows through a player. A 37-chemistry second pair on a team trying to win now is a roster hole, not a depth chart slot.

The third pair grades 23. Lundkvist is a passenger at the 35th percentile. Capobianco grades neutral with a high drive rating of 79, which says the offense flows through him without him being good enough to justify it. There is no pair on this roster the model trusts, and the third pair is among the worst marks in the league.

Then there is the crease, and it holds a surprise. Based on the model, Jake Oettinger grades 62, Strong, having stopped 0.37 goals above expected per 60 last season. Casey DeSmith grades 87, Elite, at 0.59 goals saved above expected per 60 with a 96.6 percent mark on rush chances. The model prefers the backup. That is the honest read on the numbers, and it is worth saying out loud even if the depth chart says otherwise.

Camp questions: is Heiskanen enough to carry a blue line where no pair fits? Can Harley and Myers be a real second pair, because 37 says no? And does the net become a quiet controversy, with the model grading DeSmith higher than Oettinger?

Detroit Red Wings

The lineup:

• Forwards
• L1: Emmitt Finnie - Dylan Larkin - Lucas Raymond (50)
• L2: Alex DeBrincat - Marco Kasper - Viktor Arvidsson (76)
• L3: Michael Rasmussen - Andrew Copp - Keegan Kolesar (53)
• L4: Carter Mazur - J.T. Compher - Mason Appleton (44)
• Defense
• P1: Simon Edvinsson - Moritz Seider (63)
• P2: Ben Chiarot - Justin Faulk (35)
• P3: Albert Johansson - Jacob Bernard-Docker (32)

Goalies: John Gibson (Average, 55), Daniil Tarasov (Average, 44)

As of the first day of camp, Simon Edvinsson does not have a contract. He is a restricted free agent whose camp has been waiting on a general manager the Red Wings did not have until this week, when Shawn Horcoff was named interim GM, and he plans to be at camp anyway, according to Pierre LeBrun. Say it plainly: Detroit's season is built on the Edvinsson - Seider pair, and that pair is currently unsigned. The model grades them at 63, the best number on the blue line and one of the best pairs in the division: Edvinsson at 80.4 quality with a drive rating of 88, Seider at 85.4 with a drive rating of 99. Drive rating measures how much five-on-five offense flows through a player, and 99 is about as high as it gets for a defenseman.

The surprise is the second line. DeBrincat with Kasper and Arvidsson grades 76, the best forward number on the roster, ahead of the top line's 50. The top trio of Finnie, Larkin and Raymond is the incumbent, and coaches are slow to break up what works, but chemistry grades how well three players' styles fit together, and the model's read is clear: the second line is the better line. Then the familiar Detroit problem: center depth falls off a cliff behind Larkin, and the second and third pairs grade 35 and 32. No pair the model trusts behind the top two.

Then there is the crease. Based on the model, John Gibson grades 55, Average: he stopped 0.42 goals above expected per 60 minutes but just 86.1 percent of rush chances, which is the model's way of saying he steals some and gives some back. Daniil Tarasov grades 44, also Average. An average crease behind a 63 pair is fine. An average crease behind a 35 pair is not.

Camp questions: does Edvinsson sign before the season opens, and for how long? Does the second line outscore the first by November? And can anyone on the bottom four defensemen make the model notice them?

Edmonton Oilers

The lineup:

• Forwards
• L1: Zach Hyman - Connor McDavid - Matt Savoie (86)
• L2: Ryan Nugent-Hopkins - Leon Draisaitl - Kasperi Kapanen (76)
• L3: Vasily Podkolzin - Josh Samanski - Trent Frederic (54)
• L4: Max Jones - Jason Dickinson - Isaac Howard (50)
• Defense
• P1: Mattias Ekholm - Evan Bouchard (75)
• P2: Ryan Shea - Connor Murphy (40)
• P3: Jake Walman - Ty Emberson (50)

Goalies: Frederik Andersen (Average, 46), Tristan Jarry (Below avg, 34)

Mike Babcock was hired as head coach on June 23, replacing Kris Knoblauch after the NHL cleared him to return to a bench. The model's lineup makes the big structural call: Draisaitl and McDavid separated, Hyman beside McDavid, and the two centers each driving their own line. The math likes the split well enough. The top line grades 86 with McDavid's drive rating at 100, which measures how much five-on-five offense flows through him, and Hyman at 99 behind him. The second line grades 76 with Draisaitl as an elite driver in his own right.

The new face on the top line is Matt Savoie. He played all 82 games last season at 22, scored 18 goals with 19 assists, and finished tied for the team lead in game-winning goals. The model's quality rating has him at the 48th percentile, which it calls Neutral, for a young two-way winger with strong net-front and dual-threat tags. That is not a star grade, but it is a real top-six number, and the top line's 86 says the fit works. He is not skating at camp yet, still rehabbing an offseason injury, with the team expecting him back early in the season. The projection has him on the top line when he returns.

The depth is where the numbers get thin. The third line grades 54, carried by Podkolzin, the only skater on it the model rates above passenger. The fourth line grades 50, and that grade reflects only Max Jones and Jason Dickinson. One data note: Isaac Howard has no NHL data in the model yet, so there is nothing to score on his side of the line. He tore up the AHL after a slow first look last season, and camp is his audition. After the top six, the forward group is the thinnest part of the roster.

The top pair is elite: Ekholm - Bouchard grade 75, both with drive ratings at 100. After that the defense goes quiet, with the second and third pairs grading 40 and 50. For a team built on two centers, the depth behind the top six forwards and the top pair is the question nobody has answered.

Then the goaltending, and the model is worried. Based on the model, Frederik Andersen grades 46.4, average, and Tristan Jarry grades 33.6, below average. Andersen saves a small number of goals above expected per 60 minutes; Jarry gives a few back. An average starter and a below-average backup is not a championship crease, and Andersen is not on the ice at camp yet either, so Jarry opens as the de facto starter.

Camp questions: how long does the McDavid - Draisaitl split last if the power play slumps? When Savoie returns, does the top line stick? Can Howard win a real job in camp, or does the fourth line need a veteran? And when Andersen is actually on the ice, is he enough, or is the crease the move the front office still has to make?

Florida Panthers

The lineup:

• Forwards
• L1: Brady Tkachuk - Aleksander Barkov - Sam Reinhart (81)
• L2: Sandis Vilmanis - Sam Bennett - Matthew Tkachuk (68)
• L3: Eetu Luostarinen - Anton Lundell - Carter Verhaeghe (63)
• L4: Jonah Gadjovich - Lars Eller - Garnet Hathaway (54)
• Defense
• P1: Gustav Forsling - Aaron Ekblad (56)
• P2: Niko Mikkola - Seth Jones (54)
• P3: Dmitry Kulikov - Radko Gudas (34)

Goalies: Jacob Markstrom (Below avg, 32), Akira Schmid (Average, 49)

Brady Tkachuk is a Florida Panther. The June 21 trade cost the No. 9 and No. 25 picks in the 2026 draft, a conditional 2029 first and a 2027 second, and it puts him with Barkov and Reinhart on a top line that grades 81 chemistry, which measures how well three players' styles fit together. All three are elite drivers. Brady's drive rating is 100, which measures how much five-on-five offense flows through a player, and he scored at nearly a point a game in 60 games last season. If the question is whether Florida got better at the top of the lineup, the answer is yes. Whether Brady fixed what actually broke last season is a different conversation.

What broke was not the top of the lineup. Florida missed the playoffs after back-to-back Cups because the roster got hurt, and the crease never recovered. The injuries have mostly healed: the team is reportedly healthy entering camp, with the exception of Marchand, who had offseason hip and sports hernia surgery and could miss the start of the season, which is why Vilmanis is getting a look on the second line. The crease has not healed. Based on the model, Jacob Markstrom grades 32, below average, across 44 starts, with Akira Schmid at 49 behind him. Markstrom stopped 86.6 percent of rush chances. That is a red flag on a team that just went all-in, and no winger fixes a crease.

The defense is the other unanswered question. Forsling grades 86.8, an elite driver, but Ekblad grades 60.2, neutral, and the pair sits at 56. The second pair is Mikkola, a strong driver at 64.2, carrying Jones, who grades 37.2 as a passenger, at 54. The third pair grades 34. And on team speed: Brady makes Florida heavier, not faster. The pace on this roster lives with Verhaeghe, Luostarinen and Lundell, whose third line grades 63. If the bet was that Brady solves the blue line or the net, the model says the bet is wrong. He solves the finishing. The finishing was never the problem.

Camp questions: can Vilmanis hold a top-nine job until Marchand is back? Can Mikkola keep carrying the second pair, because Jones at 37.2 is the number that decides the season? And if Markstrom repeats last season, did the Brady trade fix the wrong problem?

Los Angeles Kings

The lineup:

• Forwards
• L1: Artemi Panarin - Alex Laferriere - Adrian Kempe (76)
• L2: Kevin Fiala - Quinton Byfield - Trevor Moore (76)
• L3: Corey Perry - Erik Haula - Mats Zuccarello (54)
• L4: Joel Armia - Alex Turcotte - Scott Laughton (61)
• Defense
• P1: Mikey Anderson - Drew Doughty (67)
• P2: Joel Edmundson - Brandt Clarke (45)
• P3: Brian Dumoulin - Scott Perunovich (40)

Goalies: Darcy Kuemper (Average, 50), Anton Forsberg (Strong, 62)

Anze Kopitar retired after 20 seasons, and his absence is the whole story of this forward group. The high-end center position is no longer strong, and the question camp has to answer is whether the wings can carry it anyway. The numbers say maybe, and the maybe is loud. Fiala grades as an elite driver with a drive rating of 94, Byfield grades elite too, and Moore's drive rating sits at 89. The second line grades 76, and so does the first, where Panarin skates on an $11 million extension after arriving from the Rangers at last season's deadline. The top six grade 76 apiece. That is real wing depth, though the third line is where the veteran experiment lives.

The wrinkle is at center: in the model's lineup, Laferriere is a winger taking the draws on the top line, which is the camp experiment, and Byfield is the best natural center on the roster. The veteran insurance is Erik Haula, signed for two years at $3.6 million after 81 games in Nashville. He centers the third line between Perry and Zuccarello, and the model is blunt about the downgrade: the line grades 54 with him instead of 79 with Turcotte. Haula opens camp as the third-line center. The model prefers Turcotte in the spot, 79 to 54, so this is the place where camp has to prove the projection wrong. If Laferriere cannot hold the top-line spot, the whole structure slides down a chair.

On defense, the new captain is the engine. Drew Doughty was named the 15th captain in franchise history this week, and his drive rating sits at 85, which measures how much five-on-five offense flows through a defenseman. Brandt Clarke's drive rating is 90, the highest on the blue line. When the Kings score this year, expect the puck to start moving off these two sticks. The problem is behind the top pair: Edmundson - Clarke grade 45 and the third pair grades 40, and Clarke cannot play 60 minutes.

The crease needs a bounceback, and the numbers say what that has to look like. Based on the model, Darcy Kuemper grades 50, average, while backup Anton Forsberg grades 62, Strong, with better goals-saved-above-expected marks. A bounceback for Kuemper does not mean average; it means clearly above average, because the backup's numbers show the bar is reachable. He saved a modest 0.21 goals above expected per 60 last season. The team needs that number to roughly double.

Camp questions: can Laferriere actually play center, or does Byfield move up by November? Does Haula hold the third-line job over Turcotte? Do Doughty's minutes have to grow to cover for the bottom pairs? And what does Kuemper's bounceback look like, concretely, because the math says Forsberg is coming for the job.

Minnesota Wild

The lineup:

• Forwards
• L1: Kirill Kaprizov - Ryan Hartman - Maksim Shabanov (64)
• L2: Matt Boldy - Joel Eriksson Ek - Blake Coleman (74)
• L3: Marcus Foligno - Danila Yurov - Yakov Trenin (53)
• L4: Bobby Brink - Michael McCarron - Nick Foligno (57)
• Defense
• P1: Quinn Hughes - Brock Faber (70)
• P2: Jonas Brodin - Jared Spurgeon (68)
• P3: Olli Maatta - Zach Bogosian (46)

Goalies: Filip Gustavsson (Strong, 66), Jesper Wallstedt (Strong, 62)

Mats Zuccarello signed a one-year deal with the Los Angeles Kings on July 1, ending seven seasons in Minnesota. He scored 118 goals with 271 assists in 452 games as a member of the Wild, and for most of that run he was stapled to Kirill Kaprizov. Vladimir Tarasenko and Marcus Johansson left too. Three of the team's top seven goal scorers from last season are gone, 53 goals out the door, and the hardest one to replace is the playmaker who made Kaprizov's line work.

The camp answer, per The Athletic's reporting, is Maxim Shabanov beside Kaprizov. Shabanov is 25, a KHL star who put up 67 points in 65 games for Traktor Chelyabinsk in 2024-25, then managed 18 points in 44 games as an Islanders rookie and was not tendered a qualifying offer. Minnesota signed him July 2 for $1.6 million. The Athletic's reporting has him getting a look with Kaprizov in camp. The model has no NHL tags for him yet and grades him a passenger at the 29th percentile, which is what a blank page looks like. He might be a find. He is not Zuccarello.

The center is the bigger problem. Ryan Hartman is the top-line center in the model lineup, and the camp reporting agrees he is the favorite, but he is not a prototypical top-line center. He is a strong driver with an elite lane-clogger tag, a third-line player keeping a seat warm. The model grades the line 64 out of 100 on its chemistry scale, which measures how well three players' styles fit together.

The model's verdict is blunt: the best line on the roster is the second one. Boldy, an elite driver with elite secondary-scorer and green-light tags, with Eriksson Ek and Coleman, grades 74, ten points better than the nominal top line. The first line wears second-line sweaters. Kaprizov, who scored 45 goals with 89 points last season, is the best player on the team, and the model cannot find him a center.

The defense is the separator. Hughes, with an elite power-play quarterback tag, and Faber grade 70 as the top pair; Brodin and Spurgeon grade 68 behind them. That is a balanced top four, not a top-heavy one. Then the floor falls out: the third pair grades 46, the third line 53, and the fourth line centers on McCarron, a passenger at the 31st percentile.

Then there is the crease. Based on the model, Filip Gustavsson grades 66, Strong, having stopped 0.30 goals above expected per 60 minutes and turned away 97.6 percent of the rush chances he faced. Jesper Wallstedt grades 62, Strong, at 0.44 goals saved above expected per 60. Solid, no drama, no controversy.

Camp questions: does Shabanov stick beside Kaprizov, or does Boldy come up? Is the second line actually the first line all season? And is a trade for a real top center coming, because Hartman is keeping a seat warm?

Montreal Canadiens

The lineup:

• Forwards
• L1: Cole Caufield - Nick Suzuki - Juraj Slafkovský (91)
• L2: Ivan Demidov - Phillip Danault - Oliver Kapanen (74)
• L3: Chris Kreider - Zachary Bolduc - Alex Newhook (72)
• L4: Alexandre Texier - Jake Evans - Josh Anderson (48)
• Defense
• P1: Lane Hutson - Noah Dobson (77)
• P2: Mike Matheson - Alexandre Carrier (57)
• P3: Arber Xhekaj - Jayden Struble (36)

Goalies: Jakub Dobes (Strong, 72), Samuel Montembeault (Average, 42)

Cole Caufield scored 51 goals last season. All of them count, and 40 of them came at even strength, which is the number that matters here: the model's ratings are built on five-on-five play, and 40 even-strength goals is the engine of this roster. He added 37 assists for 88 points and won the Lady Byng, the first Canadien to reach 50 goals since Stephane Richer in 1989-90. Suzuki centered him all year and finished with 101 points, the first 100-point Hab since Mats Naslund, plus a Selke. The top line grades 91 chemistry, which measures how well three players' styles fit together, and it is the best top-line mark in the division. Last spring it carried Montreal past Tampa and Buffalo, both in seven games, to the conference final. The first line is not the question.

The question is everything behind Suzuki at center. In the model's lineup, Danault, brought back from Los Angeles in December, centers the second line with Demidov and Kapanen at 74 chemistry. Danault grades 59.8 quality with a drive rating of 88, which measures how much five-on-five offense flows through a player: he is a legitimate second-line center by the model's numbers, and his defensive reputation is real. In the model's lineup, Bolduc centers the third line at 52.4 quality, neutral. Evans is the fourth-line center. So the verdict on center depth: Suzuki is elite, Danault is a real second-line center by five-on-five chance flow, and after that it gets thin. If Bolduc is a genuine NHL third-line center, Montreal has three lines down the middle. If he is not, the second line carries a heavy load and the third line is a holding pattern.

The blue line is the roster's engine. Hutson with Dobson grades 77, the division's best pair, with Hutson's drive rating at 94. The third pair grades 36, which will not survive April.

Then there is the crease. Based on the model, Jakub Dobes grades 72, Strong, after stopping 0.58 goals above expected per 60 minutes in 43 games. Montembeault grades 42, Average. Dobes is the starter by the numbers, and Fowler grades Strong in 17 games behind them, which is the future knocking.

Camp questions: how long before Demidov forces his way onto the top line? Is the Hutson - Dobson pair the best in the East? And is Bolduc really the third-line center, or does the bottom six get shuffled by December?

Nashville Predators

The lineup:

• Forwards
• L1: Steven Stamkos - Ryan O'Reilly - Mavrik Bourque (77)
• L2: Filip Forsberg - Ross Colton - Matthew Wood (70)
• L3: Adam Edstrom - Jack Drury - Jonathan Marchessault (62)
• L4: Nils Hoglander - Alexander Kerfoot - Ozzy Wiesblatt (39)
• Defense
• P1: Roman Josi - Brady Skjei (62)
• P2: Nicolas Hague - Justin Barron (41)
• P3: Adam Wilsby - Ilya Lyubushkin (40)

Goalies: Juuse Saros (Average, 47), Justus Annunen (Strong, 71)

The model's first line puts Stamkos with O'Reilly and Bourque, and the math likes it: 77 out of 100 on the chemistry scale, which measures how well three players' styles fit together. Stamkos carries an elite secondary-scorer tag. O'Reilly brings an elite finisher tag and a strong faceoff-specialist tag. Bourque is the interesting one: 24 years old, acquired from Dallas, 20 goals and 21 assists last season, signed to a six-year extension, and an elite driver with a drive rating of 95, which measures how much five-on-five offense flows through a player. He scored 14 high-danger goals last season, in the 87th percentile among all scorers. Behind them, Forsberg's elite green-light tag drives the second unit, and Colton, with an elite play-driver tag and a drive rating of 98, is the model's favorite at the 82nd percentile, doing his damage from the second line. The forwards are new, and they are fine.

The defense is the question, and it deserves the stress test. Josi with Skjei grades 62 as the top pair. Josi is the best defenseman on the roster by quality, at the 65th percentile with strong puck-mover, point-cannon, and power-play quarterback tags. Skjei carries an elite heavy-lifting tag but grades neutral at the 43rd percentile. The model only half-believes in the pair.

Then the drop. Hague and Barron grade 41 as the second pair, and both grade as passengers, at the 33rd and 37th percentiles. Wilsby and Lyubushkin grade 40 as the third pair; Lyubushkin is a passenger at the 22nd percentile with no tags at all. No pair outside the top one clears 45. What that means in practice: Josi cannot play 30 minutes a night, and when the matchups get hard there is nowhere to hide. Two of the three pairs grade like third pairs on a good team.

Then there is the crease. Based on the model, Juuse Saros grades 47, Average, having stopped 0.06 goals above expected per 60 minutes, which is roughly break-even. Justus Annunen grades 71, Strong, at 0.30 goals saved above expected per 60 with a 96.5 percent mark on rush chances. The backup grades higher than the starter. That is what the numbers say.

Camp questions: is the new top line a first line or a famous-names line? Does Annunen push Saros for the net? And who plays the hard minutes when the second and third pairs grade 41 and 40?

New Jersey Devils

The lineup:

• Forwards
• L1: Jesper Bratt - Jack Hughes - Luke Evangelista (78)
• L2: Timo Meier - Nico Hischier - Dawson Mercer (62)
• L3: Arseny Gritsyuk - Evan Rodrigues - Connor Brown (65)
• L4: Nick Bjugstad - Cody Glass - Stefan Noesen (30)
• Defense
• P1: Jonas Siegenthaler - Dougie Hamilton (56)
• P2: Luke Hughes - Brett Pesce (60)
• P3: Declan Chisholm - Johnathan Kovacevic (36)

Goalies: Jake Allen (Average, 58.1), David Rittich (Strong, 61.9)

The model's lineup puts Luke Evangelista on the first line, at right wing, which is his natural side. The Devils paid a conditional 2028 first-round pick and a 2028 second-round pick to get him from Nashville on September 1, then signed him to a five-year, $36.25 million extension two weeks later. He is 24 and coming off career highs of 12 goals and 44 assists in 81 games. His line with Jack Hughes and Jesper Bratt grades 78 on the model's chemistry scale, which measures how well three players' styles fit together, the best mark on the roster. The model grades Evangelista as a strong driver at 69.2 quality with a drive rating of 90, which measures how much five-on-five offense flows through a player.

The stars have to stay on the ice for any of this to matter. Hughes has missed significant time in four of the last five seasons, including 18 games last year after hand surgery in November and then a lower-body injury in January. Arseny Gritsyuk is coming off upper-body surgery that ended his season in April and opens camp listed as day-to-day. The Devils finished 27th in scoring last season at 2.76 goals per game and 12 points out of the playoffs. This roster, as drawn, has two scoring lines and a lot of hope.

The hope runs into the bottom of the lineup. The fourth line grades 30 and the third defense pair grades 36. Over an 84-game season, that is where years go to die. First-year general manager Sunny Mehta added Anthony Mantha, who scored 33 goals last season, plus Evan Rodrigues and Connor Brown to fill out the forwards. The model is not convinced yet. The third line of Gritsyuk, Rodrigues, and Brown grades 65, which is fine, and then it falls off a cliff.

Jacob Markstrom was traded to Florida, so the crease belongs to Jake Allen and David Rittich. Based on the model, Allen grades 58.1, Average, and Rittich grades 61.9, Strong. Allen stopped nearly half a goal above expected per 60 minutes last season, which is quietly good work in 37 games. The model reads this as a competent tandem, not a difference-making one, which is a fair description of a team whose real question is whether its best players play.

Camp questions: does Evangelista look like a $7.25 million top-liner by Christmas? Can Hughes play 70-plus games? Is there enough in the bottom six and the third pair to survive 84 games?

New York Islanders

The lineup:

• Forwards
• L1: Simon Holmstrom - Bo Horvat - Cole Eiserman (53)
• L2: Kyle Palmieri - Jean-Gabriel Pageau - Matias Maccelli (40)
• L3: Emil Heineman - Casey Cizikas - Ondrej Palat (53)
• L4: Kyle MacLean - Brayden Schenn - Anthony Duclair (45)
• Defense
• P1: Matthew Schaefer - Ryan Pulock (62)
• P2: Adam Pelech - Tony DeAngelo (65)
• P3: Alexander Romanov - Scott Mayfield (36)

Goalies: Ilya Sorokin (Elite, 91.3), Vitek Vanecek (Below avg, 33.4)

Start with the one sure thing. Based on the model, Ilya Sorokin grades 91.3, Elite, the best goaltending mark in the division. He stopped 0.64 goals above expected per 60 minutes last season and turned away 98.2 percent of the rush chances he faced, which is the kind of number that wins games a roster has no business winning. Vitek Vanecek grades 33.4, below average, but that hardly matters. This is Sorokin's net, and the floor of the whole season sits in it.

Everything else is a retool in progress. Mathieu Darche is the new general manager. Anders Lee left in free agency, Bo Horvat is the new captain, and the big deadline swing was Brayden Schenn from St. Louis. And the team's best forward is not on the ice: Mathew Barzal, who led the team with 72 points in 81 games, is out of camp after aggravating a prior injury, with no guarantee he is ready for opening night.

The future is Matthew Schaefer, and it is already here. The No. 1 pick in 2025 won the Calder Trophy unanimously with 23 goals and 59 points, played all 82 games at 24:41 a night, and the model grades him a strong driver at 76.2 quality with a drive rating of 80, which measures how much five-on-five offense flows through a defenseman. His ceiling this season: another leap toward 25 goals and 70 points, and a real argument that he is already one of the ten best defensemen in the league. His pair with Ryan Pulock grades 62 on the model's chemistry scale, which measures how well two players' styles fit together, and Pelech with DeAngelo grades 65 behind them.

The forwards are the problem. No line grades above 53. The top line is Horvat between Simon Holmstrom and Cole Eiserman, and the model has no NHL data on Eiserman yet, so that 53 is really a grade on the other two. Then there is Schenn: a natural center, a former captain, a $6.5 million player through 2028, shown here on the fourth line at left wing. He has played wing before, but this looks like the wrong use of him, and camp should sort out whether it is.

Camp questions: is Barzal ready for opening night? Where does Schenn actually play? Does Schaefer take the next step from Calder winner to the Norris conversation?

New York Rangers

The lineup:

• Forwards
• L1: Gabe Perreault - Mika Zibanejad - Alexis Lafrenière (52)
• L2: Pavel Dorofeyev - J.T. Miller - Will Cuylle (65)
• L3: Tye Kartye - Joe Veleno - Jaroslav Chmelar (53)
• L4: Taylor Raddysh - Noah Laba - Oliver Bjorkstrand (44)
• Defense
• P1: Vladislav Gavrikov - Adam Fox (76)
• P2: Matthew Robertson - Sean Durzi (49)
• P3: Marcus Pettersson - Urho Vaakanainen (30)

Goalies: Igor Shesterkin (Strong, 79.3), Joonas Korpisalo (Average, 54.0)

Start with the facts. The Rangers finished last in the Eastern Conference at 34-39-9 and missed the playoffs for the second straight year. Then they traded for Pavel Dorofeyev and signed him for seven years and $77 million on June 30. Dorofeyev is 25, scored 37 goals last season, which would have led the Rangers by three, and added 12 more in 22 playoff games for a Vegas team that reached the Final. He is a right wing entering his prime, and the model grades him an elite driver at 77.9 quality with a drive rating of 98, which measures how much five-on-five offense flows through a player. That is where the new scoring comes from: the net, from everywhere, at volume.

Alexis Lafreniere is the other bet. He tied his career high with 57 points last season, then put up 25 points in his final 25 games after the Panarin trade pushed him into a top-line role. The model grades him a strong driver at 65.4. For the first time in his career, the projection has him opening the season on the top line at even strength and the first power-play unit. If the second half was the real player, the Rangers have two 30-goal wingers.

The catch is the top line itself. Perreault with Zibanejad and Lafreniere grades 52 on the model's chemistry scale, which measures how well three players' styles fit together, the weakest top-line fit in the division. Perreault grades as a passenger at 32.4. The second line, Dorofeyev with J.T. Miller and Will Cuylle, grades 65 and might be the better unit by November.

On the playoff question, with Fox playing like the Norris player he is: based on the model, Adam Fox grades 88.4, Elite, with a drive rating of 98, which measures how much five-on-five offense flows through a defenseman, and his pair with Vladislav Gavrikov grades 76, the surest thing on the roster. The crease is not the problem either: Igor Shesterkin grades 79.3, Strong, and he stopped 0.61 goals above expected per 60 last season. A Norris-level Fox, a 37-goal Dorofeyev, and a real Lafreniere leap is a playoff formula. The model says the skaters around them are average, and the bottom two defense pairs grade 49 and 30. The floor is higher than last place. The ceiling needs the kids.

Camp questions: does Dorofeyev start on the second line or force his way up? Is Lafreniere's second half the real player? Can a Fox-Gavrikov pair and Shesterkin drag an average roster into the playoffs?

Ottawa Senators

The lineup:

• Forwards
• L1: William Eklund - Tim Stützle - Drake Batherson (78)
• L2: Claude Giroux - Shane Pinto - Michael Amadio (68)
• L3: Warren Foegele - Dylan Cozens - Fabian Zetterlund (62)
• L4: Andre Burakovsky - Ridly Greig - Nick Cousins (50)
• Defense
• P1: Jake Sanderson - Artem Zub (72)
• P2: Thomas Chabot - Jordan Spence (61)
• P3: Tyler Kleven - Carter Yakemchuk (43)

Goalies: Linus Ullmark (Average, 45), Samuel Ersson (Below avg, 21)

Brady Tkachuk played eight seasons in Ottawa, wore the C for five of them, and put up 463 points in 572 games. Last season he scored at nearly a point a game and assisted on 11 of Dylan Cozens' goals, more than any other teammate. That is the hole: not just a left winger, but the player the offense ran through and the identity of the room. The June 21 trade sent him to Florida for two first-round picks, a conditional 2029 first and a 2027 second. GM Steve Staios said the picks are in play for immediate help. Two days later he flipped one of them.

The flip was William Eklund, acquired from San Jose on June 23 for the No. 9 pick, and the model's lineup puts him straight onto the top line with Stützle and Batherson at 78 chemistry, which measures how well three players' styles fit together. Eklund had 53 points, 15 goals and 38 assists, in 78 games last season and is signed at $5.6 million through 2028-29. He replaces the left-wing minutes and some of the playmaking. He does not replace the physicality or the captaincy. That is the honest math of the hole: no one player fills it, so Ottawa spread it around. Burakovsky, acquired from Chicago for a sixth-round pick, is a two-time 20-goal scorer who slots in as depth scoring on the fourth line. Giroux re-signed for one year and drops to the second line, where the model still grades him a strong driver. Stützle, an elite driver, is now the player the top line runs through, with Batherson, also an elite driver, on the other wing.

The defense did not lose anyone. Sanderson with Zub grades 72: Sanderson at 94.2 quality with elite defensive-anchor tags, Zub a strong driver at 80.0. Chabot with Spence grades 61, with Chabot at 87.8 quality. The top four is the best part of this roster, and it has to be, because of what comes next.

Then there is the crease, and this is the second hole. Based on the model, Linus Ullmark grades 45, Average: he stopped 0.09 goals above expected per 60 minutes, which is break-even, an average season in every sense. Samuel Ersson, acquired from Toronto, grades 21, below average, after allowing 0.36 goals above expected per 60. Ottawa lost its captain and did not upgrade its goaltending. The defense is excellent. The net is a question.

Camp questions: does Eklund look like a first-liner from day one? Who brings the snarl now that the captain is in Florida? And does Staios move one of the remaining picks for a bigger answer before opening night?

Philadelphia Flyers

The lineup:

• Forwards
• L1: Tyson Foerster - Trevor Zegras - Travis Konecny (66)
• L2: Matvei Michkov - Sean Couturier - Owen Tippett (76)
• L3: Denver Barkey - Noah Cates - Porter Martone
• L4: Alex Bump - Christian Dvorak - Nikita Grebenkin (51)
• Defense
• P1: Travis Sanheim - Rasmus Ristolainen (70)
• P2: Cam York - Jamie Drysdale (66)
• P3: Nick Seeler - David Jiricek (24)

Goalies: Dan Vladar (Strong, 60.2), Joseph Woll (Average, 49.0)

Porter Martone is the storyline. The 19-year-old, the No. 6 pick in 2025, debuted on March 31 and put up 10 points in nine games, then five more in 10 playoff games, becoming the first teenager ever to score the game-winning goal in each of his first two playoff games. He helped drag Philadelphia to the playoffs for the first time since 2020. The likeliest look has him opening camp on the third line with Denver Barkey and Noah Cates. One data note: Martone has no NHL data in the model yet, so his line carries no chemistry score.

The best line on the roster is clearer. Matvei Michkov between Sean Couturier and Owen Tippett grades 76 on the model's chemistry scale, which measures how well three players' styles fit together. Tippett is the headliner: an elite driver at 85.4 quality with a drive rating of 87, which measures how much five-on-five offense flows through a player. The concern is down the middle. Trevor Zegras, between Tyson Foerster and Travis Konecny on the top line, grades Neutral at 54.3, and the Flyers spent the summer swinging for a true No. 1 center, including the offer sheet for Leo Carlsson that Anaheim matched. Behind Zegras it is Couturier, Cates, and Christian Dvorak: honest players, none of them a top-line center.

The goaltending should be better than last season. Dan Vladar was the breakout: 29 wins, a 2.42 goals-against average, then a .922 save percentage in the playoffs, and a new long-term extension. Based on the model, he grades 60.2, Strong. The change is behind him: Samuel Ersson, who was dismal early last season, is gone to Toronto, and Joseph Woll is in. Woll grades 49.0, Average, but his .899 save percentage last season came behind a Toronto team that bled chances, and two years ago he was a .909 goalie on a division winner. If Vladar repeats last season and Woll is even league average, the crease is not the reason they would miss. They made it last year with Ersson in the room.

Camp questions: is Martone a top-six winger by December, or is he one already? Does anyone seize the No. 1 center job, or is it another deadline search? Can the third pair, grading 24, be hidden for 84 games?

Pittsburgh Penguins

The lineup:

• Forwards
• L1: Rickard Rakell - Sidney Crosby - Bryan Rust (77)
• L2: Egor Chinakhov - Evgeni Malkin - Justin Brazeau (75)
• L3: Ville Koivunen - Ben Kindel - Nicholas Robertson (72)
• L4: Connor Dewar - Tommy Novak - Hendrix Lapierre (64)
• Defense
• P1: Samuel Girard - Erik Karlsson (76)
• P2: Ilya Solovyov - Kris Letang (48)
• P3: Trevor van Riemsdyk - Kaedan Korczak (53)

Goalies: Arturs Silovs (Below avg, 12.2), Sergei Murashov (Average, 42.9)

Pittsburgh is trying to do two things at once: stay competitive while the kids arrive. The top of the lineup says they can. Rickard Rakell with Sidney Crosby and Bryan Rust grades 77 on the model's chemistry scale, which measures how well three players' styles fit together, and the second line, Egor Chinakhov with Evgeni Malkin and Justin Brazeau, grades 75 right behind it. Crosby opens his 20th season as captain, and the model still grades him a strong driver at 72.4 quality. Chinakhov grades Neutral at 75.5, which in plain terms means the model sees a good player who has not yet become a driver, and Malkin grades Neutral at 63.7.

Kris Letang is not on the bottom pair in the model's lineup. In the projection he skates on the second pair with Ilya Solovyov, and that unit grades 48. The model grades Letang Neutral at 42.8 now, which is what the age curve looks like in numbers, and Solovyov's drive rating of 92, which measures how much five-on-five offense flows through a defenseman, is doing the heavy lifting there. The best pair is the first: Samuel Girard, an elite driver at 81.8, with Erik Karlsson at 76.

The retool half is real too. The third line is all kids: Ville Koivunen, Ben Kindel, Nicholas Robertson, and it grades 72, the third-best forward unit on the roster. Every line on this team grades 64 or better, which only Carolina matches in the division.

Then the crease, which is the problem. Based on the model, Arturs Silovs grades 12.2, below average, and Sergei Murashov grades 42.9, Average. Silovs played 39 games last season with an .888 save percentage, and Murashov has five NHL games. The honest read: a team trying to compete cannot get below-average goaltending over 84 games and make the playoffs. If neither goalie takes a step, the retool wins the argument by default.

Camp questions: does Chinakhov take the next step and push for top-line minutes? How many minutes can the old core play over 84 games? Is there a goalie move coming, because this tandem is not a playoff tandem?

San Jose Sharks

The lineup:

• Forwards
• L1: Mason Marchment - Macklin Celebrini - Will Smith (86)
• L2: Adam Gaudette - Alexander Wennberg - Tyler Toffoli (55)
• L3: Barclay Goodrow - Michael Misa - Collin Graf (59)
• L4: Kiefer Sherwood - Zack Ostapchuk - Ty Dellandrea (69)
• Defense
• P1: Darnell Nurse - Jacob Trouba (65)
• P2: Dmitry Orlov - Michael Kesselring (52)
• P3: Sam Dickinson - Luca Cagnoni (45)

Goalies: Yaroslav Askarov (Below avg, 33), Alex Nedeljkovic (Average, 40)

The top line is exactly what it should be: Celebrini and Smith on the same line, with Marchment's complementary play-driving alongside them. It grades 86, the best chemistry mark of any Pacific team, and Celebrini's own numbers back it up: an 84.5 quality rating in the Strong driver tier. When the two best young players on the roster fit together this well on paper, you play them together and figure out the rest later.

Ivar Stenberg is the name missing from the lineup above. The No. 2 overall pick signed his entry-level contract in July, put up 33 points in 43 SHL games as an 18-year-old, and is expected to open the season in San Jose — he skated on Misa's wing at the Rookie Faceoff. His arrival is the real pressure point on the bottom six: the third line grades 59 and the fourth 69, and a Stenberg breakthrough in camp pushes one of those incumbents out of the opening-night picture.

The rest is a cliff. The second line grades 55, the third 59, and the fourth 69, and the average quality of the forwards falls apart after the top unit. Michael Misa centering the third line is a development story, not a winning one. There is no second scoring line on this roster right now.

The defense deserves a hard look. The top pair is the summer's veteran rebuild made manifest: Nurse and Trouba grade 65, and both carry strong drive numbers. Behind them the pairs do not fit. Orlov has a drive rating of 96, which measures how much five-on-five offense flows through a defenseman, and Kesselring matches him at 96, yet the pair grades 52. The third pair grades 45. Two good offensive defensemen who do not defend together is a problem camp has to solve.

And the goaltending is the deeper problem. Based on the model, Yaroslav Askarov grades 33, below average, giving back 0.17 goals above expected per 60 minutes, and Alex Nedeljkovic grades 40, average. A defense this unsettled needs a goalie who steals games. The model does not see one here.

Camp questions: does the Celebrini - Smith line stay together all season, or do the coaches protect the kids by splitting them? Who actually plays on the second and third pairs when the numbers fit this poorly? Where does Stenberg slot in, and whose roster spot does he take? And is Askarov ready to be a starter, because the defense in front of him is not doing him any favors?

Seattle Kraken

The lineup:

• Forwards
• L1: Bobby McMann - Matty Beniers - Jordan Eberle (65)
• L2: Mackie Samoskevich - Chandler Stephenson - Kaapo Kakko (80)
• L3: Ryan Winterton - Shane Wright - Jared McCann (47)
• L4: Berkly Catton - Frederick Gaudreau - Jacob Melanson (47)
• Defense
• P1: Vince Dunn - Adam Larsson (49)
• P2: Ryker Evans - Brandon Montour (44)
• P3: Joshua Mahura - Cale Fleury (39)

Goalies: Joey Daccord (Average, 48), Philipp Grubauer (Below avg, 35)

The model's lineup hands Beniers the first-line center job, finally, anchoring the top line between McMann and Eberle. But walk through the units and the average-team case starts to look kind. The top line grades 65. The model's best unit is the second line: Samoskevich, with a drive rating of 96, plays with Stephenson and Kakko at 80 chemistry. The third and fourth lines grade 47 each. No pair clears 49. The verdict: no skater on this roster grades in the model's Elite tier. Samoskevich and McCann are strong drivers, and everyone else is Neutral or a passenger. This is not the most average team in the league. It is a slightly below-average team with no top-end talent anywhere to lift it.

That is the crux. Teams win in this league on the strength of their best players, and Seattle's best players are good second-liners. Beniers grading Neutral as a first-line center is not a criticism of Beniers; it is a description of the roster around him. The defense is the same story: Dunn - Larsson grade 49 after years together, Montour's pair grades 44, and the third pair sits at 39. Average-ish everywhere, elite nowhere.

The crease follows the theme. Based on the model, Joey Daccord grades 48, average, saving a modest 0.18 goals above expected per 60 minutes but giving some of it back on high-danger chances. Philipp Grubauer grades 35, below average, and bleeds on the dangerous chances too. There is no game-stealer here either.

Camp questions: does the second line become the first line, because 80 says it might? Can Beniers carry a top line for 84 games without an elite winger beside him? And is there any path to a top-four defenseman, because the numbers say the blue line is three pairs of the same?

St. Louis Blues

The lineup:

• Forwards
• L1: Dylan Holloway - Robert Thomas - Jimmy Snuggerud (86)
• L2: Jake Neighbours - Connor McMichael - Pavel Buchnevich (79)
• L3: Nathan Walker - Mason McTavish - Jonatan Berggren (50)
• L4: Ross Johnston - Pius Suter - Jack Finley (48)
• Defense
• P1: Philip Broberg - Logan Mailloux (41)
• P2: Cam Fowler - Colton Parayko (53)
• P3: Tyler Tucker - Brandon Carlo (55)

Goalies: Joel Hofer (Strong, 67), Jordan Binnington (Below avg, 26)

St. Louis is about to find out. Thomas with Holloway and the rookie Snuggerud grades 86 out of 100 on the model's chemistry scale, which measures how well three players' styles fit together. That is one of the best marks in the league. Thomas has a drive rating of 98, which measures how much five-on-five offense flows through a player, with elite play-driver and trigger-man tags. Holloway carries an elite distributor tag. The model's lineup has Buchnevich on the second line, which grades 79, so the top six is real. Then the fall: the third line grades 50, the fourth 48. McTavish, a neutral at the 50th percentile with a drive rating of 23, centers the third line, which is either patience or a waste, and camp will decide which.

The blue line is a camp-long audition. Broberg is the best defenseman on the roster by quality, an elite driver at the 85th percentile with strong matchup-defender and suppressor tags, but his pair with Mailloux grades 41 because Mailloux is a passenger with a liability tag. Parayko, an elite driver at the 87th percentile with an elite suppressor tag, plays the second pair with Fowler at 53. Tucker, a strong driver with a drive rating of 84, and Carlo grade 55 as the third pair. No pair clears 55. The numbers say nobody has emerged yet.

Then there is the crease, and the model has a clear preference. Based on the model, Joel Hofer grades 67, Strong, having stopped 0.34 goals above expected per 60 minutes. Jordan Binnington grades 26, below average, at negative 0.28 goals saved above expected per 60. The model prefers the 46-game backup over the Stanley Cup winner, and it is not close. That is the honest read.

Camp questions: is the top line the best in the division? Is McTavish a third-liner or a top-six center? And who emerges on defense and in net, because the numbers say nobody has yet?

Tampa Bay Lightning

The lineup:

• Forwards
• L1: Brandon Hagel - Anthony Cirelli - Nikita Kucherov (100)
• L2: Jake Guentzel - Brayden Point - Gage Goncalves (67)
• L3: Ilya Mikheyev - Conor Geekie - Pontus Holmberg (56)
• L4: Jeffrey Viel - Zemgus Girgensons - Scott Sabourin (55)
• Defense
• P1: Victor Hedman - John Carlson (49)
• P2: Ryan McDonagh - Erik Cernak (68)
• P3: J.J. Moser - Charle-Edouard D'Astous (55)

Goalies: Andrei Vasilevskiy (Strong, 78), Jonas Johansson (Below avg, 25)

You do not break up 100. Hagel, Cirelli and Kucherov grade a perfect 100 in chemistry, which measures how well three players' styles fit together, and it is the only perfect mark in the league. Hagel's drive rating is 100, the maximum, which measures how much five-on-five offense flows through a player. Cirelli grades 83.6 as an elite driver. Kucherov is Kucherov. Generation, conversion, and conscience, all elite. The second line grades 67 behind Guentzel and Point, which is the kind of second line that wins divisions. Everything after that is ordinary: 56, 55.

Now the blue line. The model's lineup has Victor Hedman there, first name on the first pair with Carlson, and the honest question of camp is what the model thinks of him. The model grades Hedman at 43.4 quality, a passenger, with a drive rating of 27. Carlson grades 76.4 as a strong driver and is the engine of the pair, which sits at 49. That is below the second pair (McDonagh - Cernak at 68, the unit the model trusts most) and below the third pair (Moser - D'Astous at 55). Hedman at his best was the best defenseman on the planet. The model's verdict is that the five-on-five chance flow is not what it was. He is still Victor Hedman: the minutes, the matchup, the power play. But 43.4 is not a reputation grade. It is a number, and camp will show whether it is right. Carlson, signed this summer to replace the departed Raddysh, is the reason the pair is playable at all.

Then there is the crease. Based on the model, Andrei Vasilevskiy grades 78, Strong. He stopped 0.59 goals above expected per 60 minutes and turned away 91.3 percent of rush chances in 58 games. The best line in hockey and a Strong goalie: that is the whole case for Tampa.

Camp questions: does the perfect line stay together all year? Is the model's verdict on Hedman right? And is the third pair playable in May?

Toronto Maple Leafs

The lineup:

• Forwards
• L1: Matthew Knies - Auston Matthews - Jack Roslovic (78)
• L2: Nick Paul - John Tavares - William Nylander (61)
• L3: Easton Cowan - Teddy Blueger - Dakota Joshua (39)
• L4: Gavin McKenna - Colton Sissons - Brandon Duhaime (51)
• Defense
• P1: Jake McCabe - Darren Raddysh (69)
• P2: Morgan Rielly - Chris Tanev (51)
• P3: Emil Andrae - Oliver Ekman-Larsson (49)

Goalies: Sergei Bobrovsky (Average, 39), Anthony Stolarz (Strong, 73)

Start with the subtraction. Max Domi failed his physical and is out indefinitely after complications from offseason surgery, the GM announced this week. He was coming off a 36-point season and was not in the opening-night plans anyway, but it is one fewer middle-six option on a roster that turned over nearly two dozen players this summer.

Now the additions. The blue line is new at the top: Raddysh arrived in a sign-and-trade from Tampa on an eight-year deal, and the model's lineup puts him straight on the top pair with McCabe at 69 chemistry, which measures how well two players' styles fit together, the best defensive number on the roster. Raddysh scored 22 goals and 70 points in 73 games with Tampa last season, and his drive rating is 100, which measures how much five-on-five offense flows through a player: he is the power-play quarterback and the engine of the pair. Andrae came over from Philadelphia. Tanev has been cleared to play after core muscle surgery. Rielly is working through a viral condition and will be ramped up slowly, but is expected for opening night. That is a rebuilt top four, and the model likes the top pair.

The most important new face starts on the fourth line. Gavin McKenna, the No. 1 pick in June, had 51 points in 35 games at Penn State and 129 points in his final WHL season, and camp is his chance to earn a top-six spot. The model has him on the fourth line with Sissons and Duhaime at 51 because it has no NHL data on him yet. The camp question writes itself: if McKenna plays like a typical top rookie, he is not a fourth-liner for long, and a second line with McKenna on the left becomes the most interesting experiment in the Atlantic.

Matthews is the other question. The captain called last season frustrating, and the GM said at media day that he is full-go and 100 percent, smiling and having fun. The top line of Knies, Matthews and Roslovic grades 78, with Roslovic arriving from Edmonton after a 21-goal season. If Matthews is himself again, the top six is real. If he is not, nothing else on this roster covers for it. The third line grades 39, the weakest number on the roster outside the goalie question, and the bottom six is thin.

Then there is the crease. The Maple Leafs signed Bobrovsky to a three-year deal on July 1 to be the No. 1, reuniting him with Stolarz. Based on the model, Bobrovsky grades 39, Average, after a career-worst .877 save percentage last season. Stolarz grades 73, Strong, after stopping 0.38 goals above expected per 60 minutes and 94.1 percent of rush chances. Based on the model, Bob is not an upgrade. He is the famous name. Stolarz is the better goalie by the numbers, and the net is a camp battle whether the depth chart says so or not.

Camp questions: what changes if McKenna forces his way onto the second line in October? Can Matthews get back to himself? And is Bobrovsky an upgrade in net, or does Stolarz keep the job?

Utah Hockey Club

The lineup:

• Forwards
• L1: Clayton Keller - Logan Cooley - Dylan Guenther (87)
• L2: Michael Carcone - Nick Schmaltz - Kailer Yamamoto (81)
• L3: Lawson Crouse - Jack McBain - Barrett Hayton (70)
• L4: Anders Lee - Vincent Trocheck - Daniil But (62)
• Defense
• P1: Mikhail Sergachev - MacKenzie Weegar (60)
• P2: Nate Schmidt - Nick DeSimone (61)
• P3: John Marino - Andrew Peeke (49)

Goalies: Karel Vejmelka (Strong, 66)

The case for them as one of the league's best teams is in the forward lines, unit by unit.

Utah has the highest average forward-line chemistry in the league at 75.0. It is one of three teams where every forward line grades 62 or better. No other team in the league can say both of those things. Go down the lines and the reason is obvious: the quality never dips.

The top line grades 87 out of 100 on the chemistry scale, which measures how well three players' styles fit together. Keller is an elite driver at the 86th percentile with elite playmaker and dual-threat tags. Guenther carries elite sniper, finisher, and trigger-man tags. Cooley centers it with an elite trigger-man tag of his own.

The second line grades 81, a second line that scores like a first line. Schmaltz is an elite driver with elite two-way player, dual-threat, and secondary-scorer tags. Yamamoto is an elite driver with a drive rating of 88, which measures how much five-on-five offense flows through a player, and elite opportunist and straight-shooter tags. Carcone is a strong driver with a drive rating of 89. Only five teams in the league have a second line at 80 or better.

The third line grades 70. Hayton is an elite driver with a drive rating of 98 and an elite play-driver tag. Only five teams in the league have a third line at 70 or better. The fourth line grades 62, with But bringing an elite lane-clogger tag and a drive rating of 93. Four lines, no passengers, no drop-off. That is why the model thinks this team is one of the best in the league.

Now the honest caveat. The blue line has no pair above 61. Sergachev is a strong driver at the 82nd percentile with elite heavy-lifting and power-play quarterback tags, but Weegar grades neutral with a drive rating of 30, and the pair grades 60. Schmidt and DeSimone are both strong drivers by quality, but the fit grades 61. The third pair grades 49, because Peeke is a passenger at the 18th percentile with a drive rating of 4.

Then there is the crease. Based on the model, Karel Vejmelka grades 66, Strong, having stopped 0.38 goals above expected per 60 minutes across 64 games. Solid, not elite. He is the only Utah goalie with a model grade.

The model's verdict comes with an asterisk. The forwards say contender: four lines that all fit, the deepest group in hockey. The defense says prove it: a 60, 61, 49 blue line and a strong-but-not-elite starter. If the forwards outscore the problem, this is a very good team. If they do not, you know exactly where it breaks.

Camp questions: is the forward depth real enough to carry a thin blue line? Does anyone on defense emerge as a true top pair? And is this the year Cooley becomes the star the first line needs?

Vancouver Canucks

The lineup:

• Forwards
• L1: Jake DeBrusk - Elias Pettersson - Linus Karlsson (52)
• L2: Liam Ohgren - Marco Rossi - Brock Boeser (40)
• L3: Drew O'Connor - Filip Chytil - Brendan Gallagher (67)
• L4: Paul Cotter - Aatu Räty - Max Sasson (47)

• Defense
• P1: Zeev Buium - Filip Hronek (48)
• P2: Jamie Oleksiak - Tom Willander (24)
• P3: Luke Schenn - Elias Pettersson (34)

Goalies: Kevin Lankinen (Average, 41), Nikita Tolopilo (Average, 43)

Last season was the worst in franchise history: 25 wins, 58 points, dead last in the league by 14 points. The response was a full reset. Quinn Hughes was traded to Minnesota, the front office was replaced by GM Ryan Johnson and co-presidents Daniel and Henrik Sedin, Manny Malhotra is the new head coach, and the team is openly rebuilding. The present is bleak, so look at the future.

The future is genuinely exciting. The Canucks own the most-improved prospect pool in the league, headlined by Caleb Malhotra, drafted third overall, and Adam Novotny at 24th. Zeev Buium, the young defenseman, just signed an eight-year extension, and veterans like Paul Cotter and Brendan Gallagher were brought in to mentor the kids. The current blue line is a construction site, and the model grades it like one: the top pair grades 48, the second pair grades 24, and Buium, Willander, and the defenseman also named Elias Pettersson sit at the bottom of the league's quality percentiles. (Two different players share the name; the center's numbers are the ones that matter up top.) This is a group being built for two or three years from now, and it shows.

The forwards are a holding pattern with a few real players in it. Elias Pettersson centers the top line after leading the team with 51 points in the wreckage of last season. Boeser's name keeps surfacing in trade chatter, which is what happens to productive veterans on rebuilding teams. The third line grades 67, the best mark on the roster, which says more about the roster than the line.

The crease is thin. Based on the model, Kevin Lankinen grades 41, average, and Nikita Tolopilo grades 43, average. Both give back a few goals above expected per 60 minutes. Thatcher Demko, once the star of this crease, played only 20 games last season. Two average goalies behind a 24-chemistry second pair is how you lose a lot of games.

Camp questions: which of the kids forces their way into the top six and stays there? Is Buium ready for top-pair minutes, or is this a patient year? And when does Boeser get traded, because the rebuild says it is coming?

Vegas Golden Knights

The lineup:

• Forwards
• L1: Ivan Barbashev - Jack Eichel - Mark Stone (96)
• L2: Mitch Marner - William Karlsson - Braeden Bowman (82)
• L3: Brett Howden - Tomas Hertl - Victor Olofsson (68)
• L4: Alexander Holtz - Nic Dowd - Marc Gatcomb (30)
• Defense
• P1: Shea Theodore - Brayden McNabb (66)
• P2: Noah Hanifin - Rasmus Andersson (66)
• P3: Parker Wotherspoon - Jeremy Lauzon (40)

Goalies: Adin Hill (Below avg, 32), Carter Hart (Average, 38)

Mitch Marner is not the new guy. He arrived in a sign-and-trade from Toronto on July 1, 2025, signed an eight-year, $96 million deal, and played all of last season in Vegas. The model's lineup has him starting camp as the best second-line winger in hockey, with Karlsson and Braeden Bowman at 82 chemistry.

Now whether the team got worse, tested against the numbers. The top of this roster is not much worse. It is still elite. Barbashev - Eichel - Stone grades 96, the highest chemistry mark in the division, with all three skaters grading as elite drivers. The second line grades 82. The third line grades 68 with Hertl driving. The top four defensemen all grade well, with both pairs at 66. The talent did not go anywhere at the top.

It went out the bottom. Pavel Dorofeyev's 37 goals were traded to the Rangers. Alex Pietrangelo is out with a hip injury. Colton Sissons, Keegan Kolesar, Reilly Smith, and Brandon Saad all left. The fourth line grades 30 with all three skaters as passengers, and the third pair grades 40. A new coach, Ryan Craig, inherits a roster that is top-heavy in a way Vegas teams have never been.

And the goaltending: based on the model, it is not solid. Adin Hill grades 32, below average, giving back 0.17 goals above expected per 60 minutes and struggling on high-danger chances. Carter Hart grades 38, average, giving back the same per 60. Both goalies' numbers point the same direction: below the bar for a team that wants to contend. This is the thinnest crease on a contender in the league.

Camp questions: does Marner move up to the top line, because 96 says the case is there? Is the fourth line playable, because 30 says it is not? And is this goaltending real enough for a contender?

Washington Capitals

The lineup:

• Forwards
• L1: Tom Wilson - Pierre-Luc Dubois - Alex Tuch (91)
• L2: Alex Ovechkin - Dylan Strome - Jordan Kyrou (77)
• L3: Ethen Frank - Boone Jenner - Aliaksei Protas (62)
• L4: Ilya Protas - Justin Sourdif - Ryan Leonard
• Defense
• P1: Jakob Chychrun - Matt Roy (55)
• P2: Martin Fehérváry - Dylan McIlrath (42)
• P3: Cole Hutson - Vincent Desharnais (43)

Goalies: Logan Thompson (Strong, 81.2), Charlie Lindgren (Average, 43.0)

The lineup card says one thing and the numbers say another. Alex Ovechkin is back on a one-year deal for his 22nd season, and his line with Dylan Strome and Jordan Kyrou grades 77 on the model's chemistry scale, which measures how well three players' styles fit together. The real first line is the unit above it: Tom Wilson with Pierre-Luc Dubois and Alex Tuch at 91, the best chemistry mark of any line in the division. Wilson grades as an elite driver at 78.0 quality, Dubois at 77.1, and Tuch was one of the summer's two big adds along with Kyrou after Washington missed the playoffs by two points.

Ovechkin himself is the sideshow that matters. He is 40, he has 929 career goals, and he still led the team with 32 goals and 64 points in 82 games last season. He signed for $4.25 million against the cap to chase the playoffs, not the record.

Ilya Protas is in the NHL group at camp. The 20-year-old was the AHL's rookie of the year with 29 goals and 66 points in 69 games for Hershey, then put up four points in four NHL games at the end of last season. The model's lineup has him on the fourth line with Justin Sourdif and Ryan Leonard, which leaves Ivan Miroshnichenko as the odd man out. One data note: Ilya Protas has no NHL data in the model yet, so his line carries no chemistry score.

The blue line is the concern: all three pairs grade 55 or worse, the weakest group in the division on the model's math. The crease is the counterweight. Based on the model, Logan Thompson grades 81.2, Strong, and he stopped 0.81 goals above expected per 60 minutes last season, which is elite territory in plain words. He is signed through 2031. The plan is fairly clear: outscore the blue line's problems.

Camp questions: does the Wilson line stay together all season, or does Ovechkin's line reclaim the top spot? Where does Ilya Protas settle once the games start? Can Thompson cover for the thinnest blue line in the Metro?

Winnipeg Jets

The lineup:

• Forwards
• L1: Kyle Connor - Mark Scheifele - Gabriel Vilardi (70)
• L2: Nino Niederreiter - Vladislav Namestnikov - Cole Perfetti (31)
• L3: Cole Koepke - Adam Lowry - Alex Iafallo (46)
• L4: Nikita Chibrikov - Morgan Barron - Isak Rosen (47)
• Defense
• P1: Josh Morrissey - Dylan DeMelo (32)
• P2: Dylan Samberg - Neal Pionk (43)
• P3: Mario Ferraro - John St. Ivany (45)

Goalies: Connor Hellebuyck (Strong, 76), Stuart Skinner (Average, 43)

Connor Hellebuyck did not report to training camp and was suspended on Wednesday. He went public with his trade request on August 27, saying he had asked privately months earlier, and general manager Kevin Cheveldayoff addressed the media this morning at 11:30 ET. The reported market is thin: talks with Buffalo, interest from Utah, Carolina, and New Jersey, and one report describing the offers as terrible. The camp crease, for now, belongs to Stuart Skinner and whoever stands behind him.

The model grades goalies too, and it grades Hellebuyck. Based on the model, he is a 76, Strong, having stopped 0.43 goals above expected per 60 minutes last season and turned away 96.4 percent of the rush chances he faced. That is despite the down year: 23 wins against 23 losses with a 2.86 goals-against average and a career-worst .895 save percentage, and the Jets missing the playoffs at 35-35-12. The year before, he won the Hart Trophy and his third Vezina, then backstopped the United States to Olympic gold in February. The model still sees a strong goalie. Skinner grades 43, Average.

So how reliant is this team on its goalie? The skater base is not the worst in the league, but the numbers explain why the model is worried. The top line is legitimate: Connor and Scheifele, with Scheifele's elite playmaker and trigger-man tags, plus Vilardi, grade 70 out of 100 on the chemistry scale, which measures how well three players' styles fit together. Then the crater: the second line grades 31, a hole in the middle of the lineup, and the third and fourth lines grade 46 and 47. The real alarm is the top pair. Morrissey and DeMelo grade 32, the worst top-pair mark in the league. They are individually fine: Morrissey is a strong driver at the 78th percentile with a drive rating of 80, which measures how much five-on-five offense flows through a player, and DeMelo is a strong driver with a drive rating of 84 and a strong matchup-defender tag. They just do not fit together, and the model cannot make them fit.

The honest answer: the skaters are not bad everywhere, but they are thin in exactly the places a missing goalie exposes. A 31-chemistry second line, a 32-chemistry top pair, and an average backup in net. If Hellebuyck is not there, they are in trouble. The model says so plainly.

Camp questions: does the Hellebuyck standoff end before opening night, or does Skinner own the net? Does anyone split up Morrissey and DeMelo, because 32 is the worst mark of its kind in the league? And can the second line be fixed, because 31 is a hole in the middle of the lineup?

Build your own

Every lineup above started with the Whiteboard's lineup model. I adjusted for injuries, real positions, and what camp reporting suggests coaches will ice, then ran each one back through the Whiteboard's chemistry engine. Every score is the model's own math on a realistic lineup.

Now build yours. Open your team on the Whiteboard, put together the 20 you'd ice on opening night, and watch the chemistry number move as you shuffle the pieces. Share your lines, compare notes with other fans, and tell us where you think these 32 got it wrong. This is one version of all 32 teams. The fun is making your own.