Official 2021 NHL Season Previews

Your last minute hockey fix before the season starts

Published in

Towards Data Science

48 min readJan 12, 2021

I recently built a projection model for the 2021 NHL season. A high-level overview of this model can be found here, but the gist of it is that I used regression to determine how good each NHL player was at each thing they do and how much they would do each thing for their team, aggregated these values on the team-level to determine how well each team would do these things, and then simulated the season 10,000 times to determine the most likely outcomes for the season. Final simulations were conducted on January 6th, 2021 with updated rosters as of that date.

First and foremost, for those who like charts: Here are all teams broken down by their projected standings points and probability of making the playoffs:

And are the teams broken down by their respective divisions:

One theme you may notice here is that teams in the North Division are more likely to make the playoffs than teams outside of the North Division who are projected to earn a similar amount of points. This is because there are only 7 teams in the North Division.

North Division

These divisions are close enough to one another that the one with 7 teams is inevitably going to be the one in which it is easiest to make the playoffs, but it just so happens that the North Division is also the weakest division by team strength. This divisional weakness starts at the bottom. The average playoff cut-off for this division is 60.63 points.

Ottawa’s 13% probability of making the playoffs is not the lowest estimate of any team, but this is largely due to the structure of the Canadian Division, as Ottawa does project to be the worst team in the NHL. This should come to the surprise of nobody, but there is a contingent of voices who will disagree, citing the improvements that Ottawa made over the off-season. I vehemently disagree that Ottawa improved their team by any degree significant enough to be good, though.

Ottawa’s roster is littered with terrible depth players who struggle to drive play, and their hopes for success rest largely on the backs of three players: Matt Murray, Brady Tkachuk, and Thomas Chabot. Murray looked to be the NHL’s next young superstar goalie when he technically won two Stanley Cups as a rookie, but has fallen off a cliff since then and was one of the NHL’s worst goalies last year, which makes his projection of being a roughly average goaltender almost feel too kind to him. Brady Tkachuk has already cemented himself as one of the NHL’s best drivers of dangerous offensive scoring chances — a feat that becomes even more impressive when you consider the quality of team he’s played on while achieving this feat — but he gives some of that back defensively, and more importantly, he’s one of the worst shooters in the NHL. There’s an argument to be had that the danger of the chances that he creates is being overrated by my data, which would in turn mean that his shooting is being underrated, but either way, the conclusion is still roughly the same: He has not managed to have a significant positive impact on his team’s scoring rates. Thomas Chabot has been an excellent offensive defenseman at a very young age despite playing some of the toughest minutes in the league (in terms of ice time), but he also gives a lot back defensively, and can’t be considered a truly elite offensive defenseman until he manages to fix that or take his offense to yet another level.

Without all 3 of these players putting up incredible campaigns, it’s hard to envision a path to the playoffs for this team.

This may feel low for a team that just qualified for the 16-team playoff and then knocked off the defending Stanley Cup Champions in the play-in round. But this is a team that was quite mediocre in the regular season and lost their starting goaltender in Jacob Markstrom who many fans and media members considered their MVP. (I personally believe Elias Pettersson was their MVP, but the fact remains that Markstrom was also very valuable.) Additionally, while they won their series against St. Louis fair and square and deserve credit for doing so, it stands to reason that they were a bit fortunate to do so despite getting heavily out-shot and out-chanced, and we shouldn’t take that six-game sample where Jordan Binnington was the biggest factor to mean all that much.

I personally like the decision that Vancouver made to let Markstrom walk and the decision to replace him with Braden Holtby, whose contract presents a much smaller risk even if he does perform as poorly as I project him to, or even if he performs as poorly as he did last season. But in the short term, Holtby should be a lot worse than Markstrom was, and the fact remains that this team has major defensive problems, with most of their skaters projected to have a negative impact on the rate at which their team concedes expected goals. They’re better at offense than they are at defense, but they’re not great at that either.

This team has a decent shot at making the playoffs — they’re not dead in the water by any means — but even if we take it for a given that Elias Pettersson will put up another insane on-ice shooting percentage for the third year in a row, they’re going to need either Holtby to return to Vezina form, Demko to return to games 5–7 against Vegas form, or their entire team defense to take major strides if they want to be a Stanley Cup Contender.

Oilers fans will probably feel this projection is too low after their team finished the 2019–2020 regular season on pace for 96 points with roughly a 90% chance of making the playoffs in the traditional 16-team format were the entire season played out, but a deeper glance at last year’s performance shows where the cause is for concern. The Oilers feasted on the league’s best power play — the best the NHL has seen since the 70’s — and the league’s second-best penalty kill. Neither of these facets of the game are nearly as repeatable as even strength play, where their goal differential of -16 was over twice as far behind that of the Buffalo Sabres (-6) as it was ahead of that of the Los Angeles Kings (-20). That is not good company to be in. Moreover, future forecasts are built off of more than one season, and Edmonton has a track record of being a much weaker team than they were last year.

Just about every player on this roster is projected to be good at shooting, with Draisaitl projected to be the very best shooter in the league. They also have a few players projected to have very strong impact on the power play, so while they’re extremely unlikely to replicate last year’s mark of 29.5%, they’re also likely to be well above average on the power play. Their top two players are two of the best in the league, and although the two of them could be far better if they worked on their defensive game, the rest of Edmonton’s issues are with the rest of their lineup. Their goaltending duo combined to be roughly average last season, but they’ve both historically been worse than that and they’re now a year older. Yamamoto and Nugent-Hopkins are both good players — Yamamoto may even prove to be great in a larger sample size — but they’re well below average compared to the 3rd and 4th best forwards on other contending teams, and their forwards beyond those four are very weak. Their defense isn’t the worst in the league, but it’s well below average and projects to be worse than it could be if ice time were allocated a bit more optimally.

This team is a coin flip to make the playoffs thanks to their top two players and the division they play in, but they’re a long shot to do any sort of real damage.

It’s clear that whatever magic Winnipeg had in 2017–2018 when they made the Conference Final is long gone, along with nearly their entire defense corps, but I’m not convinced that they’re as bad as they showed last season where they barely finished the season with a higher 5-on-5 xGF% than Detroit and were carried to the bubble on the back of a tremendous, Vezina-winning and Hart-worthy season from Connor Hellebuyck. That’s not to say they’re a good team either — I’d call them average — but in the NHL’s weakest division, it would be a disappointment if they missed the playoffs.

Paul Stastny may not have scored a lot of points last season, and he doesn’t project to be a great shooter, but his underlying impact is still very strong and inserting him as the second line center should go a long way for Winnipeg’s play driving. Their bottom-6 struggles at converting on their scoring chances, but they do a decent job of generating those chances and they’re actually quite strong defensively as a unit, so I don’t consider them a problem, or at least not Winnipeg’s biggest problem. Winnipeg’s biggest problems come in the form of their defensemen that aren’t named Dylan DeMelo, and the defensive play of their top line.

If their top line can return to the level of a typical poor-defensive first line and not the absolute tire fire that it was last year and their GM can manage to snag a half-decent defenseman or two off waivers, this could be a pretty good team. If those things don’t happen, this team will have to rely on Connor Hellebuyck for another Hart Trophy level season. If they don’t get that either, they’re in big trouble.

It speaks volumes that a season where Calgary finished 8th in the Western Conference in points percentage and won their play-in series in the bubble against Winnipeg is considered a massive disappointment. This isn’t a sexy hockey team, and I don’t think anybody in Alberta is extremely excited about their Stanley Cup odds for this season. But they’re also not a bad hockey team either, grading out as roughly average in just about every facet of the game, which is about in line with how they played last year despite a down-year from most of their top players. In the worst division in the league, being an average team should be enough for the Calgary Flames to make the playoffs just under 2/3 of the time.

Goaltending has been a problem over the past few years for Calgary, but the addition of Jacob Markstrom means it will likely be a problem for their opponents going forward. This team’s top players took a step back last season, and will likely never reach the levels they peaked at in 2018–2019, but they still project to be good, and the bottom of Calgary’s lineup is generally made up of players who drive play quite well, which should give them an edge on the competition if their top line just manages to break even.

Their defensemen aren’t great, but would look a lot better if Michael Stone were replaced with somebody half decent and Oliver Kylington managed to take some steps forward. Ultimately, the ceiling for this team doesn’t look all that high, but the floor doesn’t look too low either.

My projections are not unique in that I rank Montreal significantly higher than they’ve ranked in the actual standings over the past few seasons. For the most part, everybody’s projections roughly mirror team performance from the prior seasons, and the way that performance is quantified and weighed across those seasons determines whose projections are more accurate than others, with everybody’s projections being fairly accurate. So why is everybody so much higher on Montreal than the actual standings were in the past? My suspicion is backup goaltending and a poor record in one-goal games. Montreal’s goal differential over the past two seasons is only -2, which suggests they’re more of an average team than a truly bad one. But more importantly, this number is horribly deflated by backup goaltending; they’re a +33 over that sample with Carey Price in net. It’s worth noting that most teams will look better if you only count their goal differential in games where they played their best goaltender, but the difference here in Montreal’s case is egregious. Additionally, their underlying metrics at 5-on-5 over this sample rival every team not named Vegas for second best in the NHL.

Unlike the backup goaltenders Montreal has employed over the past couple years, Jake Allen was great last year and projects to be good next year. If this can lead to Carey Price getting more rest and playing even better in the games he does start, that’ll be huge. But even outside of goaltending, the rest of this roster is quite good. Despite underwhelming point totals that are solely due to low power-play ice time and a poor performance in those minutes, the Tatar-Danault-Gallagher line has spent two full seasons facing some of the toughest competition in the NHL and obliterating them by every measure — including goals. The rest of their roster features quite a few question marks, but two of their new additions in Tyler Toffoli and Josh Anderson are both strong play drivers who have seen their point totals deflated by poor on-ice shooting, their bottom-6 drives play well, and their defense features good offensive players.

If Ben Chiarot, Joel Edmundson, and Jonathan Drouin can put up results that are in line with how good Montreal’s management perceives they are, or Claude Julien can limit their ice time, this team can make some serious noise.

Speaking of teams dragged down by backup goaltending and a poor record in one-goal games: Meet the Toronto Maple Leafs. Their goal differential with Frederik Andersen in net over the past two seasons is +72. This may be a tad juiced up by the fact that Toronto generally starts Andersen in the first half of back-to-backs and hangs their backups out to dry, but it still clearly proves that this is a strong team so long as their goaltending plays well. This 90% figure that I’ve listed above is the highest of any team, which should rightfully spark optimism and excitement among Leaf Nation. But it should be noted that this is a double-edged sword, the other edge of which states that if Toronto does miss the playoffs in this division, or even lose in one of the first two rounds, there will be no more excuses for them to make. This is a total 180 from facing the likes of Boston and Washington in the first round for three straight years; the likelihood of failure is lower here, but the humiliation and soul searching that will be born out of a failure here if it does occur will be much higher.

While this team’s defense corps has rightfully been touted as a major issue for years, the addition of T.J Brodie and emergence of Travis Dermott means the defensive play of these six should actually be above average, even with Morgan Rielly leading them in ice time. However, the team is still projected to be a hair below average defensively as a whole thanks to weakness from their forwards. But defense is not a prerequisite for success; this team’s goaltending is strong, and their offense is near the top of the league, projected to score by far the most goals (thanks in part to the weakness of their division but mostly due to their skill).

It’s a very strong roster, and one that would be even further ahead of the pack if not for the puzzling decision to add below-replacement level players in Wayne Simmonds and Zach Bogosian at seven-figure salaries and grant them a roster spot. If youth can take those spots early and be decent at worst, this team could be even better.

West Division

The West Division is top-heavy in a uniquely balanced fashion, featuring three excellent teams at the top, three weak teams at the bottom, and two mediocre teams in the middle. This Division may not be the best as a whole, but it is certainly the strongest defensively, and the smartest pick for a Division to be represented by the Stanley Cup Champions may be this one. The average playoff cut-off for this division is 62.34 points.

The Los Angeles Kings aren’t quite the worst team in the NHL, but thanks to the division they play in, they have the lowest probability to make the playoffs. Their future certainly looks bright, with many arguing that their prospect pool is the best in the NHL after drafting center Quinton Byfield. But as of right now, the probability that this team makes the playoffs is very slim; they project to be very bad offensively and defensively, near the bottom of the league in both goals for and goals against. Todd McLellan always seems to get a little bit more out of his teams than he should as far as play-driving goes, and I could therefore see this team being a bit better at driving play than I project them to be, but it’s hard to see them being even average in that regard and their goaltending and shooting should still be a major problem even if they are.

It’s tempting to brush aside these projections due to the possibility that Drew Doughty and Jonathan Quick will rebound; both were strong players as recently as 2017–2018. But projections are based on results, and the fact that both of them project to be so poor despite the use of a 3-year sample just shows how terrible they’ve hurt their team in both of the past two years.

Anze Kopitar is still going strong, but even if he maintains that and the aforementioned duo returns to a respectable level of play this season, this roster is still weak enough across the board that it’s hard to envision any kind of damage being done this year. It would also help if this team could add a decent 6th defenseman, and perhaps a 5th to replace Kurtis Macdermid, and then take some minutes off of Doughty; that might also help him be better.

The San Jose Sharks have been floated around as a bounce-back candidate for this season, and for good reason: They’ve historically been a strong team, and there’s reason to believe that last year was more of an outlier than proof of what they are. Additionally, they project to be a roughly average play-driving team, and their biggest weaknesses are their goaltending and their shooting; components which are least repeatable from year-to-year. However, the ceiling for this team isn’t very high either; their play-driving is more likely to decline than improve thanks to the age of their skaters, they’ve lost some of the talent that made them a dominant play-driving team two years ago, and almost every single one of players they’ve retained is projected to perform a lot worse than they did that year. It’s easy to get caught in the trap that this team’s upside should be roughly equal to what they did just two years ago, but the truth is that it’s a lot lower.

Contrary to popular belief, the biggest problem with the San Jose Sharks last season wasn’t actually goaltending; it was an uncharacteristically poor performance from their skaters. Doug Wilson did his best to make sure San Jose’s skaters won’t be their biggest problem next year when he let go of Aaron Dell, certainly San Jose’s best goalie last season and arguably their MVP (it was a bad year) and acquired Devan Dubnyk from Minnesota. The decision to walk away from a decent backup in Dell and add one of the worst goalies in the league in Dubnyk to a team that already has one of the worst goalies in the league in Jones makes it a near-certainty that poor goaltending will drag this team down harder than their potentially half-decent group of skaters will.

Now, any time that anybody says something about goaltending, it should be noted that goaltending is random enough that it’s not even outside a reasonable range of possibilities that one of these guys has a good season. But even outside of goaltending, this roster is fairly underwhelming, and this team will need a Vezina-caliber campaign out of one of their goalies to make any real noise.

The Anaheim Ducks project to be a half-decent defensive team, ranking a shade above average in goaltending and expected goal defense despite allowing a high shot volume. They’re built to give up a ton of shots without giving up a ton of goals and score enough on counter-attacks to win a decent number of games. This is another team that very likely does not have the potential to make any serious noise in the playoffs, but that could absolutely make the playoffs in place of a superior team over the short 56-game sample and potentially knock out another superior team in a playoff round. Their goal for this season may be to re-build, and that’s probably what’s going to happen, but they’re also well-built to be a pesky team that wins more frequently than they should.

John Gibson still projects to be above average after a rough 2019–2020 season, but one has to wonder if last year was more of a fluke after being burned out and his multiple years of being a truly elite goaltender are more of what he truly is. If that’s the case, this team stands to be even more annoying than usual. They’re terrible offensively; it’s not to see how they rank so low when you look up-and-down their roster and gaze at the poor offensive play-driving of just about every member of their roster.

But it is possible that some of their young players — namely Max Comtois, Isac Lundestrom, and Sam Steel — could all make significant improvements to their offensive play driving and raise Anaheim to a semi-respectable level as a team. It’s easy to forget about Anaheim after their play over the last two seasons, but they’re far from the worst team in the league, and the playoffs isn’t a total pipe dream.

The Arizona Coyotes profile is basically that of a rich man’s version of the Anaheim Ducks: They’re horrible at driving play offensively, they give up a ton of shots but most of them are low quality, their goaltending is elite, and their shooters are decent enough to score on counter-attacks. A rich man’s version of a bad hockey team is still a bad hockey team, but like Anaheim, this team is built to be annoying and win a bunch of games they’re not supposed to, and a slightly higher percentage of those games than Anaheim is.

On paper, this team’s goaltending is built about as ideally as you could build a goaltending tandem: Two guys who are both elite are set to split games roughly 50/50. But hockey is not played on paper, and both Raanta and Kuemper have struggled with injuries in the past, which could damage their in-game performance or more likely force a third goaltender to play in place of them if they miss time. The rest of this team’s roster is not built to handle even an average performance from their goaltenders, so if they don’t get a strong performance out of them they should be in trouble.

There’s some room for improvement among their skaters; young players like Clayton Keller and Lawson Crouse can manage to improve upon the way they’ve played in the past, and underrated #7D Ilya Lyubushkin can snag Jordan Oesterle’s lineup spot, and both of these things should be enough to push this team into a playoff spot if they get strong goaltending. But this team will live and die by their goaltending, and they’re screwed if they don’t get it.

The scorekeepers at the Xcel Energy Center where the Minnesota Wild play have shown a tendency to erroneously report that shots are taken further from the net than they actually are, which I’ve written about in detail. This leads the quality of shots they allow and take to be underestimated. But even after building a model that adjusts for this, I’ve still come to the conclusion that this is an elite defensive team with bad offense and average shooting. That sounds like the perfect recipe for a half-decent playoff team until you start to talk about their biggest weakness: goaltending. Their goaltending tandem last season was far and away the worst in the league, and if they had just received average league goaltending, they would’ve been a bonafide playoff team and Stanley Cup Contender.

Minnesota will be replacing both of their goaltenders from last season — their two biggest problems — due to Alex Stalock’s injury and Devan Dubnyk’s trade, but the guys they added don’t project to be all that much better than the guys they lost. Outside of them, though, this roster actually looks pretty good. The forwards are as good at anybody as forechecking their opponents into submission which is on display in their defensive impacts, and their blue line is arguably the best in the league. (And there won’t be an argument for anybody else if Greg Pateryn and Matt Dumba regain the decent form they’ve shown in prior years.) Nick Bonino’s analytics are a meme, but from what I can tell, Minnesota added a damn good all-around forward to their team when they got him from Nashville alongside a 2nd round pick in exchange for Luke Kunin.

Add Kirill Kaprizov to the mix, and you’re looking at a team that with the potential to be a dark horse contender for the Stanley Cup if their goaltending is half-decent, which isn’t that much of a longshot given the volatility of the position.

St. Louis followed up their 2019 Stanley Cup by leading the Western Conference in points percentage in the regular season. They did lose to Vancouver in 6 games in the bubble, but that was more of a PDO fluke than a true indicator of their ability or even their performance in those games. Normally, everything I just said there would indicate that I’m talking about a bonafide Stanley Cup Contender with >80% playoff odds that should finish with at least 100 points over an 82 game season. And this team is still damn good. But they’re notthe same team that accomplished those feats.

Much has been made of the loss of Alex Pietrangelo, and there’s nothing wrong with fans and media doing so; he was an elite defensemen, one of the very best in the league, and losing him will hurt St. Louis. But the losses of Alex Steen and Jake Allen, both of whom were very good last year, also stands to do a bit of damage to this team. Mike Hoffman and Torey Krug are both good players who stand to add a lot of value, but if St. Louis wants to extract the maximum value possible out of them, they’ll need them to gel on the power play and succeed there, which is a lot easier said than done and has the potential to go south if there isn’t enough puck to go around; they’re not actually great offensive players at even strength, which is a general area of weakness for this team.

This should still be an excellent defensive team with a strong goaltender playing most of their minutes, but if Binnington plays more like he did in the bubble or Ville Husso plays below the league average level that he’s projected to, they could find themselves in danger, since there’s always some risk to being a weak offensive team in a strong defensive division.

A reasonable expectation for the Colorado Avalanche was that they’d see a bit of a drop in the standings this season which would stem from a drop in their league-leading PDO from last season. But then Joe Sakic traded Nikita Zadorov and two second round picks for Brandon Saad and Devon Toews (in two separate trades). Now this team is one of the strongest, most balanced teams in the NHL from the top to the bottom, ranking in the top-ten in every facet of the game. Goaltending is cited as a concern for this team — one I obviously don’t share since I’ve ranked them tenth — but this team is strong enough across the board that they’d be well-equipped to manage even some of the worst goaltending in the league if that’s what they got.

Colorado’s roster is spearheaded by Nathan MacKinnon, a superstar center who’s just a bit behind being the true best player in the league, but unlike what they were two years ago, this is in no way a one-man show ran by him alone. This roster is extremely strong up and down the board; even fifth defenseman and 9th forward Valeri Nichushkin Ian Cole project to be excellent two-way play drivers. When you look at the names behind that tenth-place goaltending ranking, it’s a bit easier to understand why people are concerned about two guys who have had injury troubles and put up their results in easier roles, but I still feel that fear is misguided.

The biggest concern for this team in my eye isn’t goaltending, but rather their top-two defensemen. Samuel Girard has the best impact on penalty differential of any player in the league, and penalties being left out of this model definitely underrates him, but it’s still concerning that Colorado is projected to give the most ice time to a defenseman who’s a net negative play driver. Shifting those minutes from Girard and Johnson to Toews and Makar might make this team even more dominant.

Vegas may not win the President’s Trophy like they’re projected to, but something will have to go seriously wrong for them not to outshoot the vast majority of their opponents, and for the shots they take not to be closer to the net than the shots they allow. It’s what they’ve done for both of the past two years now after transitioning a bit from the good play-driving and elite transition-offense team they were in their first season. But for whatever reason, this approach hasn’t translated to actual goals for or goals against nearly as well as it’s translated to expected goals. They’re still comfortably positive in goal differential over each of the past two seasons, but they’re suspiciously low relative to where they rank in expected goal differential. The goals against can be explained by poor goaltending — Marc-Andre Fleury has fallen off a bit and their backups have both been terrible — but the shooting is a more complicated thing. It’s possible that expected goal models underrate the quality of chances they generate off their heavy forecheck, and thereby underrate their 21st ranked shooting. The bottom line is that this is unquestionably a very strong team, but they may fail to meet the projection of an elite team if they can’t convert on their dangerous chances.

Outside of a few players, this team is remarkably well-balanced up and down the lineup. Their worst player looks to be Cody Glass, who struggled last year in a small sample but is still young and has the potential to grow an average player, and every other player looks decent at worst. This team is built to make the playoffs even with horrible goaltending and shooting; they’ve done it for a few years now and just added a Norris Trophy caliber defenseman in Alex Pietrangelo.

But if Robin Lehner can wrestle away some of the games that Marc-Andre Fleury is projected to play and match his performance over the past couple of seasons, this team should comfortably make the playoffs and potentially win the President’s Trophy even if they struggle to convert on their scoring chances. If they can manage to convert on their scoring chances at a normal rate, this will the best team in the NHL — bar none.

Central Division

The Central Division features last year’s Stanley Cup Champions, last year’s worst team, one more bad team, and four other teams that all have a shot at the playoffs. The average playoff cut-off for this division is 62.11 points.

Detroit’s disastrous 2019–2020 campaign makes it hard to reconcile the idea that they have even a 1.5% chance at making the playoffs, let alone a 15% chance at making the playoffs and a 1% chance at winning the division. But this team’s biggest problem last season was Jimmy Howard, who is gone, and they made a number of other moves to improve their team. Nobody is saying they’re a good team — well, I’m not — but they’re certainly not as bad as they were last year and I don’t think they’re quite as bad as they’re made out to be.

Understandably lost in the shuffle from last year’s disaster season were a few positives: the strong play of the Bertuzzi-Larkin-Mantha line when all three were healthy, the strong defensive play of Patrik Nemeth, and the half-decent goaltending of Jonathan Bernier. Nemeth isn’t the most exciting name on your blue line in an ideal world, but a team who just had the worst season in decades has to take positives where they can get them.

Detroit’s slim hopes for the playoffs rest on the shoulders of improvement from players like Zadina with sky high potential but poor results at the NHL level, strong play from low-key additions like Sam Gagner, Jon Merrill, and Thomas Greiss, and their few bright spots from last season maintaining or building upon that performance. It’s probably going to be another painful year, but we’ll see.

It doesn’t really feel like the Chicago Blackhawks are trying to win this season, and I can’t really blame them. Jonathan Toews and Kirby Dach set to miss most of or all of the season, and their goaltending is projected to be the worst in the league. If those two things weren’t true, they might actually have the makings of a bubble team, but they can’t control those injuries and acquiring good goaltending is possible but easier said than done. The best course of action for this team is probably to let the season play out and then deciding at the trade deadline what long-term direction the team wants to take, which is exactly what it seems like management is content to do.

Dominik Kubalik came into the NHL at the age of 24 and tore the league apart from day one, leading the NHL in 5-on-5 goals per hour with strong underlying metrics to boot. I expect his individual shooting not to be quite as fortunate next year, but I’d be shocked if he’s not still a damn good player. He’s also not the only intriguing young piece on this roster; Boqvist and Debrincat are both good players right now and both of them have very high potential.

Unfortunately for the Blackhawks, the rest of their roster is fairly weak and their only established NHL goaltender is a guy whose track record is being the worst in the league. It’s totally possible that Subban or Collin Delia (their projected backup) put together a half-decent campaign in net, and that would go a long way for this team, but it’s gonna take a lot more for them to be legitimate contenders next season.

The Florida Panthers sure are a strange team. Every year, a bunch of people predict that this is the year they’ll finally take a step forward and become a playoff team, but it just never seems to happen. To be fair, they generally finish a lot closer to a playoff spot than last place, so it’s not like these predictions are horribly missing the mark in any given year, and it’s not like they were born out of a bad process. If somebody told me before the 2019–2020 season that they thought Sergei Bobrovsky would put together a decent season, I wouldn’t necessarily call them an idiot just because it didn’t happen. Sometimes a good guess is just wrong, and it’s fair to say this team has felt primed to take a step forward for a while now. They’re a strong offensive team with great shooting, but a poor defense and poor goaltending mean they’ve only got a 44% chance for this to be the year they take that step forward and make the playoffs.

Losing Mike Hoffman and Evgenii Dadonov hurts, but Bill Zito made numerous acquisitions in Vinnie Hinostroza, Patric Hornqvist, Carter Verhaege, and Radko Gudas that stand to benefit his team and mitigate the damage from those losses. As a whole, this should be a strong offensive team that buries their chances like nobody’s business. Sergei Bobrovsky isn’t far removed from 2017–2018 where he was arguably the best goaltender in the NHL, or even 2018–2019 where he was still solid. His strong track record means he gets the benefit of the doubt and is only projected to be a bit below average next season despite being legitimately horrible last year. Chris Driedger started only 11 games last season, but he was strong enough in those games that he’s projected to do quite well in a backup role next season.

If Driedger can steal some of Bobrovsky’s minutes and maintain that performance or Bobrovsky can bounce back to what he’s been in the past, this team has a solid chance at making the playoffs, but their defensive deficiencies will likely cut them off from being a true contender even if that does happen.

Last year’s Columbus Blue Jackets are sort of the opposite of what the Florida Panthers have been over the last few years: Everybody predicted that they would miss the playoffs after making it for countless seasons and then they made it anyway. John Tortorella drove a mediocre-on-paper roster to expected goal shares that were well above average and was rightfully rewarded with the Jack Adams Trophy. It’s possible and even likely that Tortorella’s impact is “baked” into this projection by way of a roster full of average-at-best play drivers looking notably above average. But this roster’s lack of talent shines through in the estimated impact of their goaltending and shooting, both of which are below average. Their strengths and weaknesses balance out to create a team whose chances of making the playoffs are a coin flip, which feels totally right to me.

Whenever the topic of “Seth Jones’s analytics” comes up, I always hear people say that Columbus’s system is not conducive to good advanced stats. This always perplexes me since Columbus defensemen like Savard, Gavrikov, and Kukan have historically been very good players according to the analytics. The defensive game of their forwards is also held in high regard. The goaltending projection may be difficult to reconcile given Joonas Korpisalo’s performance in the bubble, but he’s actually been a rather poor goalie over the last few regular seasons, which represent a significantly larger sample that more emphasis should be put on. Even if their goaltending improves and is league average this year, this team still has other problems. Their shooting is projected to be bottom of the barrel, and despite the Tortorella effect that I’ve touched on, their play driving isn’t projected to be all that far above average.

They’re a coin flip to make the playoffs but it doesn’t feel like this team has a ton of upside to be a true Stanley Cup Contender.

The Nashville Predators remind me a bit of the Calgary Flames in that both of them finished the season top-8 in the conference by points percentage, and both are talked about like they just had an awful year. This isn’t because these were truly awful seasons, but because more is expected out of these teams as they have generally provided it in the past. As it stands, Nashville is projected to be just a hair above average in every facet of the game outside of goaltending where they’re projected to be average. This is good enough for respectable marks of 62 projected points and a playoff probability of 56%, but it’s hard to get too excited about the Stanley Cup chances of an aging roster that isn’t elite at any one thing.

I mentioned earlier that the Minnesota Wild might have the best blue line in the NHL. The reason I was forced to use the word might is because the Nashville Predators blue line also exists. Led by reigning Norris Trophy Winner Roman Josi and backed by two defensemen who are both arguably better than he is, this is an excellent blue line that is in contention for the best and could snag that spot with ease if Dante Fabbro can figure out his terrible offensive impact. Unfortunately, the forwards are a bit below average due to the loss of underrated bottom-6ers Craig Smith and Nick Bonino who crushed easy minutes last season alongside Rocco Grimaldi.

This is another team where most of their roster has reached or passed the age where they traditionally peak, so it’s hard to see from where they’d take another massive step and become contenders. But at worst, they should still be in the mix for a playoff spot — comfortably so if Juuse Saros can take more than 53% of the starts.

The Dallas Stars made the Stanley Cup Finals last year, how do they only have a 60% chance of making the playoffs this year? Well, the answer to that is two-fold: Their roster has been ravaged by injuries and more importantly, they were never all that good to begin with. Fancy stats aside, they finished the regular season with a goal differential of +4 and they finished the playoffs (excluding round robin games) with a goal differential of +1. There’s something to be said for edging out one-goal games and breaking even in the playoffs where the competition is harder on average, but these are still not the marks of an elite team. It’s fair to say that some of their poor results were due to unlucky shooting last season, but that was roughly balanced out by elite goaltending and there’s no reason to believe that this team would have even average shooting results if luck were truly “normalized” as they’re projected to be 28th next season. This is definitely a good team, and that’s reflected in their 60% probability of making the playoffs in a good division despite injuries, but they’re not nearly as good as you’d be led to believe if all you heard was that they made the Cup Finals last year.

Whether it’s due to the system they play or it’s due to the talent these players possess, there’s an impressive number of players on this roster whose defensive impact is projected to be good next season. But outside of their top three forwards and top defenseman, the number of players on this roster whose offensive impact is projected to be a significant positive is practically non-existent. Part of this can be blamed on the absence of Tyler Seguin, who’d be projected to play big minutes and drive offense well if not for his surgery that’s set to keep him out for the entire season. But either way, this team is not good offensively, though they could likely mitigate those woes if they altered their system a bit in order to have their forwards take more of their shots, as their defensemen are much worse finishers than forwards.

Lastly, although it has been a strength over the past few years and is projected to be a strength here, goaltending is a large question mark. Anton Khudobin is projected to be very good, which only makes sense because he’s been very good over the past few years, but his good play has always come in a backup role. He’s a small 34 year old goalie, and it’s tough to say how he will do in a starting role, much less how Jake Oettinger will do when he sees his first taste of NHL action. If these guys do struggle, this team could be on the outside looking in.

After years of the analytics community screaming at the top of their lungs that the Carolina Hurricanes were primed to become a good hockey team, they finally did so in 2018–2019 and proved it was no fluke in 2019–2020. Now it’s time for this team to take the next step and prove they’re not just a good hockey team, but a great one. They’re only likely to reach those heights if they can improve upon their shooting, which is currently projected to be the worst in the league. There’s an argument to be made that like Vegas, they’re not actually this bad at shooting (or this good at offense) because expected goal models and raw shot counts overrate the quality of the shots they take. That might be the case, but either way this is a legitimately good team who can still generate actual goals well enough to have just over a 2-in-3 chance of making the playoffs. They’ll also be able to avoid their kryptonite, the Boston Bruins, for at least two rounds if they do make it, which bodes well for them.

This chart doesn’t add any new information that wasn’t clear from the last chart; this team has decent goaltending, excellent offensive play-driving, decent defensive play-driving, and horrible shooting. Even maligned players like Jake Gardiner and Nino Niederreiter who didn’t score a lot of points last year are still projected to be above average play drivers. Ironically, a big part of the reason this team projects to be so terrible at shooting is new players in Brady Skjei and Vincent Trocheck who have historically been quite poor at it. They’ve got some holdovers who are also terrible at it, like Jordan Staal, Brock McGinn, and Jordan Martinook, but it’s quite interesting that their two new, big additions both only serve to make them worse at a facet of the game in which they already struggle. Shooting is quite random, so perhaps it will change for these two in particular, or perhaps it’ll only become even worse once Carolina’s system takes a grip on their sticks.

The goaltending is the most interesting part of this roster, because Reimer had historically been quite poor prior to this year and then was excellent in limited minutes. He still projects to be roughly average, but if he reverts to a poor level of play this team could be in trouble. On the flip side, if Reimer maintains his strong play and this team finally scores actual goals at a similar rate to expected goals, they are a legitimate Stanley Cup Contender. They might be one anyway.

How many teams in the NHL could lose their best player for the entirety of the season and still have a 41% chance of making the playoffs? Maybe ten? Now consider that the Tampa Bay Lightning don’t have a 41% chance of making the playoffs; they have a 41% chance of winning their division and an 87% chance of making the playoffs. The 2019–2020 Stanley Cup Champions are on an island of their own as far as team strength when everybody is healthy, and they’ve got a legitimate argument for the best team even without their best player. They’re elite defensively, their goaltending has finally put together results that almost match their reputation, and their offense doesn’t generate a ton of quantity but more than makes up for it with quality and shooting.

Andrei Vasilevskiy has been touted as one of the league’s top goaltenders for years now, but only recently have the analytics supported a conclusion anywhere near that one. Now is what matters, though, and he and McElhinney are projected to come together to make a top-5 goaltending duo. This roster’s biggest strength lies in its shooting talent, with Brayden Point and Steven Stamkos leading the charge as two of the league’s best shooters. There’s a part of me that wonders if those two and the rest of this roster will struggle to score as far above expectation without Nikita Kucherov buttering their bread, but even if they fail to match what they’ve done in the past they should still be very good. Just as impressive is this team’s defense; up and down the lineup they are littered with excellent defensive players. If they just managed to sign somebody a bit better than Luke Schenn and trade Tyler Johnson for your typical replacement level forward who isn’t Tyler Johnson, defensively, they’d rival Minnesota for the best defensive team in the NHL. As it stands, they’re not quite there but they’ll probably end up allowing fewer goals anyway thanks to the strength of their goaltending.

A lot of different things could happen to this team, but it’s very hard to envision a season where they’re not one of the best teams in the league.

East Division

Last but certainly not least — actually best — is the East Division. This is the NHL’s strongest division, and it’s perfectly designed to create a bloodbath that will keep a few young, hungry teams desperate for playoff experience out of the hunt. The average playoff cut-off for this division is 62.09 points.

Make no mistake about it: This is a bad hockey team. But it’s not a terrible hockey team, and more importantly it’s a young team with a lot of room to grow. A 20% chance at making the playoffs is enough incentive for this team to put up a fight,

As I mentioned in the introduction, this roster includes Corey Crawford who will not play and excludes Sami Vatanen who will play. I imagine incorporating these players would slightly change things and harm the team but not to an extreme degree. Regardless, this team should feature strong special teams if their players can maintain their performance, but at even strength their offense should be poor and their defense should be terrible due to numerous defensive drags like Pavel Zacha, Miles Wood, P.K. Subban, Connor Carrick, Nikita Gusev, and Jack Hughes. Hughes in particular had an interesting rookie season; he was one of the worst shooters in the league and was horrible defensively, but actually drove offense rather well. He wasn’t an elite generator of offense, but even doing well at the age of 18 in an undersized frame bodes well for his development long term. However, we are not working with the super long term here, and next year he is projected to perform the same as he did this year, being a strong offense generator but a heavy net negative due to his poor shot and defensive play.

Overall, this isn’t a sexy roster, but if Mackenzie Blackwood can match last year’s performance in a legitimate starter role, they could make some noise.

The hype for the Buffalo Sabres this season doesn’t match a team projected to finish only one point ahead of the New Jersey Devils, and I believe that’s because the hype is built on seemingly plausible scenarios while the projection is built on what has actually happened in the past. Perhaps Taylor Hall could return to his Hart Trophy form while Rasmus Dahlin takes a step forward and becomes an above average defenseman at the NHL level, but both of those things could also not happen and this team really stands to struggle if that’s the case, which it isn’t currently projected to be. Additionally, this team added some players who are projected to cause major harm to their team.

New addition Taylor Hall is projected to do a lot of good for this team, which is awesome. But new additions Cody Eakin and Tobias Rieder are projected to do about as much damage. I’d be a lot more bullish on this team if those roster spots were just handed to unproven gambles like Josh Ho-Sang or Arttu Ruotsolainen, but handing those spots to these players is a concern. This defense is also really ugly, with five of six defensemen projected to have a negative impact. If that projection is in line with reality, it’s really going to harm the goal-scoring output of their forwards. Speaking of the forwrds, it’s quite concerning how few of them provide any sort of defensive value. Most of them are well below average. These kind of play-driving weaknesses are the type that you can stomach if they are offset by strong shooting or play-driving, but in this team’s case, neither of those things are projected to come.

There are a few players I could see taking massive steps: Taylor Hall, Rasmus Dahlin, Linus Ullmark, and even Rasmus Ristolainen, just to name a few. Hockey is weird enough that Ristolainen could totally translate his physical tools to decent on-ice results. But unless one, or preferably a few, of these major steps forward is actually taken, this team is going to be in an uphill battle for a playoff spot.

Why are the Rangers projected to finish closer to dead last in the Division than a playoff spot? Team defense. It’s very hard to be the worst defensive team in the NHL and still make the playoffs. You might argue that the Rangers did it last year, but I would counter by saying they weren’t quite the worst last year, they weren’t quite in a playoff spot when the season was shutdown, they got swept in the play-ins, and most importantly they got a lot worse when they lost Jesper Fast and added Jack Johnson. And that is despite the fact that they had incredibly fortunate seasons from Artemi Panarin and Mika Zibanejad that neither player is likely to replicate going forward. On the flip side, defense is the easiest area to improve in, and they did show some signs of improvement there late in the season. There’s plenty of room for optimism with this team, but it should probably be cautious.

I understand that the Rangers have flirted with healthy scratching Jack Johnson, sparking excitement among their fanbase, but based on the way that every NHL team ever has handled veteran defensemen they pay seven-figure salaries to, I’m very skeptical that they’ll actually do that — at least to start the season. If I’m wrong and Johnson never plays a game, they should improve considerably; he’s currently projected to play big minutes and do big damage in them. There’s also the near-inevitability that Kaapo Kakko will improve on his disastrous rookie season, but that stands to be balanced out by Adam Fox crashing down from his own excellent rookie season that completely defied expectations in the opposite direction. It’s also possible that Shesterkin puts up an elite rookie campaign and starts a larger percentage of the games than projected, and that Alexis Lafreniere who does not even have a projection winds up putting together an elite Calder Campaign.

But with this team, every time I find something to be optimistic about, I blink and find something beside it to be pessimistic about, like the fact that Zibanejad and Panarin are a lot more likely to fall below their projected shooting than they are to fall above it. There’s a lot of ways that things could go for this team, and they’re going to make some people look very stupid no matter what. But fans can take solace in the fact that this team announced their intention to rebuild not long ago, and they’re already well ahead of schedule.

This New York Islanders team became the enemy of the analytics community two years ago when they defied expectations and made the playoffs two years in a row, winning one round the first time and two rounds the second time. Virtually nobody expected them to make it the first time, and very few expected them to repeat the second time, but things have changed now and the analytics have caught on to the fact that this is a strong defensive team that can score a few goals here and there. It’s hard to see the upside for this team being any higher than the fortunate run to the Conference Finals they just enjoyed, and even that sounds like an extremely high bar to hold them to, but it’s a guarantee that this team will be a pain in the ass to play on a nightly basis.

Say what you will about the length of the contracts that Lou Lamoriello has handed to his bottom-6 forwards, but as of right now they’re still very strong defensive players who forecheck their opponents into submission on a nightly basis. The loss of Devon Toews is one that I believe will really hurt this team, especially if a big chunk of his minutes are absorbed by Nick Leddy, but outside of Leddy this defense is quite deep and capable in their own end. The top-6 also features strong shooters, so if they can adjust their system to make their forwards take a larger percentage of their shots and Varlamov can play as well as projected (or better yet, as well as he did last year in his first year under Mitch Korn), this team is going to win a bunch of games where they get outshot.

Ultimately, though, this strategy doesn’t tend to succeed against top teams in the playoffs; it didn’t succeed against Tampa and it didn’t succeed against Carolina. If the Islanders want to win the Stanley Cup, they’ll need to figure out a way to drive play a bit better.

The Washington Capitals are not as good as the teams that won back-to-back President’s Trophies and then a Stanley Cup the year after. In some ways, they’re actually better. They’re mostly worse, as their goaltending projects to be below average and their shooting projects to be only 5th. (That’s low by their standards.) But this team has a better chance of out-shooting and taking higher quality shots than their opposition, which, according to public expected goal models, is something they did not do during the later years of the Trotz era. The shift from the style of play that made them so successful was probably made more out of necessity than it was a conscious decision, and they’ve ultimately become worse, but it’s interesting to see them profiled as an above average play-driving team.

Alex Ovechkin is still an elite shooter and a very good generator of offense at even strength and on the power play. Just don’t ask him (or Evgeny Kuznetsov or John Carlson or Justin Schultz) about playing defense at even strength, and chances are you’ll be talking to a pretty solid defensive player. This team is full of them, and with the number of elite penalty killers they feature, they’d probably be up there with Minnesota if they weren’t dragged down by these four playing big minutes. But they’re still a strong defensive team as a whole. Their weakness is in net, and that’s likely to be mitigated by Ilya Samsonov playing a larger percentage of their games than he’s projected to. The loss of Henrik Lundqvist should hurt either way, though, as Craig Anderson is very unlikely to be even half-decent at his age.

This team has their strengths and weaknesses. They’re hanging on to a dream of contention that they’ve had for about fifteen straight years and achieved once. Considering that, they’re actually built fairly well — a 56% chance at making the playoffs in the league’s toughest division is no joke.

The Philadelphia Flyers live to frustrate data scientists. Every other year they make the playoffs with around 95–100 points, and the year after that they miss with around 85–90. At some point, you’d think a smart person would just say “odd year, Flyers are going to suck” and rob odds makers around the world, but I’m not confident enough to do that just yet. Instead, I’ve estimated that the Flyers have a 60% probability of making the playoffs and should finish with an average of 62 points. They’re a strangely built team in that they’re weak at generating shot quantity but great at generating shot quality, and inversely, they’re weak at suppressing shot quality but great at generating shot quantity. Overall, they look just a bit above average offensively and defensively, and in a tough division that’s only good enough for a 60% chance at making the playoffs.

Right off the bat, it’s clear to me that Ivan Provorov should not be the quarterback on the #1 power play. Perhaps that’s what Erik Gustafsson was signed for, and if so, he can really help out there — but not enough to offset his rough impact at even strength. Another lineup decision that could improve this team is giving Carter Hart a larger percentage of their starts than they have in the past; Brian Elliott isn’t a horrible backup but he’s definitely not good and 58% is not enough for a young star like Hart. Even outside of those lineup decisions, there’s more room for improvement here in players like Joel Farabee, Morgan Frost, Philippe Myers who could follow Ivan Provorov and Travis Konecny’s path of posting poor results early on and then improving later on. There’s also room for a rebound season for Shayne Gostisbehere, but I think it’s more likely that he winds up a healthy scratch and eventually traded as he does not seem to fit into Alain Vigneault’s system.

On the flip side of things to be optimistic about, this is the Philadelphia Flyers in an odd-numbered year, and they’ve no longer got the new coach bump going strong from Alain Vigneault. Flyers fans have every right to be cautiously optimistic about their team, especially after what they showed in last year’s regular season, but this team is not a guarantee for success by any stretch.

The Pittsburgh Penguins suffered an embarrassing loss to the Montreal Canadiens in the play-in round. But they were 7th in the league in points percentage, which means they wouldn’t even have played a play-in round if they were in the Western Conference, and more importantly, weird stuff happens in 4-game sample sizes. They had elite underlying metrics last season before half of their team got injured, and strong underlying metrics even after suffering these injuries. They’re a damn good hockey team that got rid of their biggest problem in the off-season, and while their top players are no spring chickens, they’re still elite players and nothing less than contention should be expected out of this team.

You may notice that I spoke a tad higher of the Penguins than the numbers probably suggest I should. That is entirely because of their goaltender. Tristan Jarry was fine in 2017–2018 and 2019–2020, but he was atrocious in the two games he played in 2018–2019. This really drags down his projection — a lot more than it should. The fact that this has happened suggests that I should probably weight my regressions based on ice time or games played in the future. I didn’t make any arbitrary decisions to change anybody’s projection, but I tested doing so for Jarry and removing 2018–2019 from his sample, and the Penguins jumped by about three points on average. My official projection for the Penguins is 64 points, but if they wind up with 67 I reserve the right to say I told you so.

Outside of Jarry, though, you can see what makes this lineup so good. I’m not a fan of some of their new additions in Kasperi Kapanen, Cody Ceci, and Mike Matheson, but I also absolutely love all three of those players if you stand them next to Jack Johnson, and I also love Jason Zucker. I think this was a good team dragged down by injuries last year, and I think they got better. But ultimately, they won’t live or die by the new guys; they’ll live or die by Guentzel, Crosby, Malkin, Marino, Letang, and Dumoulin; all of whom I project to be excellent.

A lot of people seem surprised that the Boston Bruins have the highest probability of winning their division of any team, but that’s what happens when you win the President’s Trophy and then bring a better team to the table than the one that won it. And yes, that is what Boston did. Last year they featured elite goaltending, elite team defense, an elite power play, strong shooting, but a rather mediocre offense at even strength. This year, they’ve made additions to improve this, and done so successfully enough that their biggest weakness of the past is an area where they now rank ninth in the division.

It may seem like a hot take to say that the Bruins got better, but I’m comfortable defending it. Forwards are more important than defensemen, and Ondrej Kase and Craig Smith are better hockey players than Zdeno Chara and Torey Krug were. The biggest concern regarding the players they lost is that the power play will struggle without Torey Krug, and although my regression analysis suggests that Krug was actually not all that important to the power play, I still think that concern is totally legitimate. Power plays in hockey aren’t as simple as just plugging in a player’s RAPM xGF/60 and generating offense at exactly that rate, and I could see the performance of Pastrnak and Bergeron being weighed down by the loss of Krug more than regression analysis suggests.

With that being said, this team can stand to take a hit or two in certain facets of the game and still kick ass. They’re just too well built not to. Their defense may look shallow, but it’s manned by arguably the best defenseman in the league and supported by an underrated gem in Grzelyck whose role is bound to grow. It would take their league-best goaltending falling off a cliff, some atrocious shooting luck, and probably a few injuries in a true catastrophe season for this team to miss the playoffs. Meanwhile, virtually nothing unlikely needs to go right to make this team a bonafide Stanley Cup Contender.

Conclusion

As the purists always say, hockey isn’t played on a spreadsheet. (Or more accurately, on a Dell in R.) With COVID-19 running wild in the United States and the timeline for a vaccine just out of reach for this season, there’s no guarantee that a season will even be completed and played out. Hell, these simulations were run using a schedule that has already had a few games postponed.

This all just adds more uncertainty to these projections. Part of the beauty of hockey is that even the most robust, granular projections can be totally disproven by force of sheer luck. Even if you respect and believe in my projections, you shouldn’t stop watching your favorite team until they’re mathematically eliminated regardless of what I say.

I hope you enjoyed reading this. And regardless of whether you agree or disagree with my projections and the methodology which I use to create them, let’s all hope for a fun, healthy 2021 NHL season for everybody!

Official 2021 NHL Season Previews

Your last minute hockey fix before the season starts

North Division

West Division

Central Division

East Division

North Division

West Division

Central Division

East Division

Conclusion

Written by Patrick Bacon