I simulated the 2020-2021 NHL regular season a million times to estimate what is likely to happen. (I used a computer to help me.) To estimate the probability of the home team winning each game I used my prediction model, Magnus. Curious readers will find lots of detail following that link, but, very briefly:
The method I used is strongly similar to the one I used last year, with some key improvements; some of the explanation is copied from last year's preview.
Using the isolated abilities of their players and their coach, I can form an estimate of how each team will peform "in a vaccum", that is, before considering the schedule of games which determined opponents and fatigue. As is my usual habit, positive values mean "more shots" or "more goals" as appropriate, and thus on offence red (more than average) is desirable and on defence blue (fewer than average) is desired. Similarly, positive shooting impact is good and negative impact is preferable for goaltending; the axis for the goaltending distributions are flipped so that good appears to the right for both shooting and goaltending.
North | West | Central | East |
---|---|---|---|
CGY![]() |
ANA![]() |
CAR![]() |
BOS![]() |
EDM![]() |
ARI![]() |
CHI![]() |
BUF![]() |
MTL![]() |
COL![]() |
CBJ![]() |
N.J![]() |
OTT![]() |
L.A![]() |
DAL![]() |
NYI![]() |
TOR![]() |
MIN![]() |
DET![]() |
NYR![]() |
VAN![]() |
S.J![]() |
FLA![]() |
PHI![]() |
WPG![]() |
STL![]() |
NSH![]() |
PIT![]() |
VGK![]() |
T.B![]() |
WSH![]() |
Using the above, I can make some summary graphics for the league as a whole.
The most basic result in the sport is the shot. To win a team will shoot often and from dangerous locations and with some measure of shooting skill. First, we look at even-strength shot rates, weighted by historical shooting percentages from the given locations. This gives an idea of "total offence" or "total defence", before we consider shooting talent and goaltending talent. I call this estimate expected goals, or xG for short. The zero point is set to what we saw in 2019-2020, around 2.5 goals per hour of 5v5 play. If the league opens up it's possible for this year's average to be above zero (or below zero if the league becomes more defensive). The "NHL" button indicates the average of the team rosters as constructed currently, as is typical for before in-season injuries the expected strength is a little better than last year's average, especially on defence.
The Maple Leafs and Golden Knights have the strongest offence creation, and Boston, Columbus, and Tampa Bay have the strongest 5v5 defences.
The same measurement above, but for special teams: power-play offence on the x-axis, and penalty-kill defence on the y-axis. Power-play defence and penalty-kill offence do not especially interest me for now.
This season most of the extreme values are weaknesses—the Wild power-play looks to be extremely bad and Detroit special teams are likely to be generally dire, as they were last season.
Expected penalty differentials for teams are computed from expected icetimes for each player, multiplied by individual tendencies to cause their team to take or draw penalties, not merely individual rates. The inability of the Islanders to draw penalties is a mystery to me, among many others. Detroit's heavy tendancy to both draw and take penalties is especially unfortunate for them considering their special teams.
Of unusual importance in this year's model (also every other year) is the impact of goaltending, and, conversely, shooting talent. The coefficients for goaltender and shooter ability are not as easily interpretable as the previous measurements, but they can still be understood with odds ratios; for instance, the odds of a shot taken against Ottawa (who parted ways with their old, weak goaltender only to acquire a young, weak goaltender by trade) becoming a goal are 5% higher than a similar shot taken against another team's goaltenders. In Winnipeg, the high value is primarily due to Connor Hellebuyck (the best goaltender in the league) unlike in Arizona and Dallas, where both goalies in the expected tandem are very strong.
As expected, the variation in goaltender talent (which rests on the shoulders of, at most, a handful of people per team) is much larger than the spread of shooting talent, which is derived from many skaters.
All of the above is taken entirely from the team's roster and coach, without regard to the schedule. Of course, not every team has the same schedule, which affects how well they are likely to perform. Not all teams are equally affected by fatigue. The strongest effects from rest are seen when teams play after playing the night before. The table below shows how many times each team plays 'tired' in this sense, as well as how many times they play against a team that is tired. This season, with much less travel, contains a lot more back-to-back play but much tighter balance; with the bulk of those games being played between two teams who both played the previous night.
The most important factor of the schedule, though, is not rest but the fact that teams do not all play against the same opposition. To measure this, only simulation will suffice, the results of which are presented below.
Team | Mean points | Standard Deviation |
---|---|---|
Leafs | 66.8 | 7.0 |
Canadiens | 63.1 | 7.0 |
Jets | 62.4 | 7.1 |
Flames | 61.4 | 7.1 |
Oilers | 60.7 | 7.1 |
Canucks | 58.0 | 7.1 |
Senators | 55.5 | 7.1 |
Each team's bar is centred on the average point total obtained in the simulations for each team, sorted with the highest averages to the right, with the divisions indicated by colour. The changing colour intensities indicate "stanines", that is, each coloured square shows half of one standard deviation.
Thus, we expect around six of the teams to finish in the darkest box, around ten to fall into the adjacent boxes, around seven to fall into the next pair of boxes, around four into the next pair, and one or two teams to fall into the faintest boxes. These one or two teams, which I make no attempt to predict ahead of time, will be much discussed. The proximate cause of their success or failure will doubtless be a superhuman goaltending performance, like Carey Price in 2014-2015, or a horrific cavalcade of injuries, like Columbus in 2014-2015, or an exuberance of last-minute goals and hot shooting, like Calgary in 2014-2015. Part of why I make predictive models is that I enjoy knowing just how unlikely are the various unlikely things that happen every year.
Playoff cutoff: 59.7 points
Team | Playoff Chance |
---|---|
Leafs | 83% |
Canadiens | 68% |
Jets | 65% |
Flames | 59% |
Oilers | 56% |
Canucks | 41% |
Senators | 28% |
Every team's most likely finishing position is marked. Toronto are most likely to finish first in the division (the marked 35%) but have a small possibility of missing the playoffs altogether. Vancouver are markedly weak (especially with the departure of their starting goaltender, who covered a multitude of flaws last year) and Ottawa weaker still, the four "middle" teams are not sharply separated in expected results although stylistically they're all quite distinct.
Team | Mean points | Standard Deviation |
---|---|---|
Golden Knights | 66.6 | 7.0 |
Avalanche | 64.5 | 7.0 |
Wild | 63.2 | 7.0 |
Blues | 62.1 | 7.0 |
Coyotes | 61.6 | 7.1 |
Ducks | 59.8 | 7.1 |
Kings | 58.5 | 7.1 |
Sharks | 54.2 | 7.1 |
Team | Playoff Chance | |
---|---|---|
Golden Knights | 76% | |
Avalanche | 67% | |
Wild | 60% | |
Blues | 54% | |
Coyotes | 51% | |
Ducks | 41% | |
Kings | 34% | |
Sharks | 16% |
The three California teams are the weakest, especially San Jose who correctly identified their most pressing need in the off-season and then addressed it by making their goaltending tandem even weaker. Vegas and Colorado are the class of the division; neither one has any weaknesses to speak of and Vegas has an extremely strong offence at both 5v4 and 5v5.
Playoff cutoff: 61.4 points.
Team | Mean points | Standard Deviation |
---|---|---|
Hurricanes | 66.8 | 7.0 |
Lightning | 66.1 | 7.0 |
Blue Jackets | 63.3 | 7.0 |
Stars | 63.3 | 7.0 |
Predators | 61.3 | 7.1 |
Panthers | 60.8 | 7.1 |
Wings | 54.9 | 7.1 |
Chicago | 53.8 | 7.1 |
Team | Playoff Chance |
---|---|
Hurricanes | 78% |
Lightning | 74% |
Blue Jackets | 60% |
Stars | 60% |
Predators | 49% |
Panthers | 46% |
Wings | 18% |
Chicago | 15% |
The central falls into four neat pairs: the great (Tampa and Carolina), the good (Dallas and Columbus), the average (Nashville and Florida), and the bad (Chicago and Detroit). Carolina is still constructed broadly along "analytics darling" lines, with overwhelming offensive volume and below-average shooting talent, but defensive strength and an above-average goaltending tandem makes them the class of the division. Tampa are more "well-rounded", either a little bit or a lot better than average in every aspect, even after losing their best player (Kucherov) to injury.
Playoff cutoff: 61.5 points.
Team | Mean points | Standard Deviation |
---|---|---|
Bruins | 69.5 | 6.9 |
Penguins | 65.2 | 7.0 |
Flyers | 61.1 | 7.1 |
Capitals | 61.1 | 7.1 |
Rangers | 60.9 | 7.1 |
Islanders | 58.8 | 7.1 |
Sabres | 56.9 | 7.1 |
Devils | 56.4 | 7.1 |
Team | Playoff Chance |
---|---|
Bruins | 88% |
Penguins | 72% |
Flyers | 50% |
Capitals | 50% |
Rangers | 49% |
Islanders | 37% |
Sabres | 28% |
Devils | 26% |
The east is the most top-heavy division, with Boston heads and shoulders above Pittsburgh who are in turn markedly stronger than the rest of the division. Boston's strengths are well-known; primarily an extremely strong 5v5 defence and an elite powerplay (both expected to be league-best) as well as goaltending near the top of the league. Pittsburgh benefits here from a very strong head coach in Mike Sullivan, who gets credit for the improvements of many players putting up better results upon moving to Pittsburgh.
Playoff cutoff: 61.1 points.
Since the league still hasn't yet implemented Gold drafting (which guarantees a full slate of exciting games for fans of all teams, win or lose, while eliminating tanking and giving the best picks to the worst teams), many teams will once again play many games at the end of the year which would be in their best interests to lose. Although there will be many trades and confusions between now and the end of the year, these are the early probabilities (including all three lotteries) for who will get the 2021 first overall draft pick.
Team(s) | Chance of First overall pick |
---|---|
SEA | 11.5% |
CHI | 7.2% |
S.J | 6.9% |
DET | 6.4% |
OTT | 6.0% |
N.J | 5.3% |
BUF | 5.0% |
VAN | 4.2% |
L.A | 3.9% |
NYI | 3.8% |
ANA | 3.1% |
MIN* | 2.7% |
FLA | 2.7% |
NYR | 2.7% |
EDM | 2.6% |
PHI | 2.6% |
WSH | 2.6% |
NSH | 2.4% |
CGY | 2.3% |
ARI** | 2.3% |
STL | 2.1% |
WPG | 1.9% |
MTL | 1.8% |
CBJ | 1.7% |
DAL | 1.6% |
COL | 1.2% |
T.B | 0.9% |
TOR | 0.8% |
CAR | 0.8% |
VGK | 0.8% |
BOS | 0.4% |
PIT* | 0.0% |
The Pittsburgh chance is zero because their first round pick for this draft was traded to Minnesota; their chance is higher than it would otherwise be for this reason. Arizona does not have a first-round pick in this draft at all, but I do not know how their 2.3% of probability mass will be distributed to the other teams.