Teams with the best goal differential in NHL history

[Big Phil]

It might surprise you a bit. Or not. But either way, Tampa in 2019 is a surprising 35th on this list at +103. The 1977 Habs - no shock - lead this list with 216 and it isn't even close. Boston in 1971 is next with +192. Montreal in 1978 follows with +178. Actually, 5 of the top 6 teams are Habs teams.

Tampa, despite their huge year in 2019 isn't as high as you might think. Ottawa in 2006 is ahead of them, so is Chicago (no kidding) in 1974. As are the 1976 Islanders and the 1980 Sabres. Good teams, but still ones that surprise me that they are ahead of Tampa from this year.

Most of the top 40 are from the 1970s and 1980s based on the fact scoring was higher than it is today. Before Tampa in 2019 the highest was Detroit in 1996 with +144. Only the 1944, 1945 and 1959 Habs had at least +100 in goal differential pre-expansion. Two of them are war years, but 1959 is even more impressive since it was a 70 game schedule and teams played each other 14 times a year with lots of familiarity. The 1996 Wings are the only team in the 1990s with at least +100, which surprised me. The 2001 Devils are the only team in the 2000s after the 2006 Sens.

I am just thinking of other dominant teams that could have been here. The 1930 Bruins pro-rated definitely make this list but played just 44 games yet were +81. The teams that just miss the +100 cut are the 1993 Pens at +99. The 2006 Red Wings were +96. It surprised me that the 2002 Wings weren't even close.


Team Game Finder | Hockey-Reference.com

 

[DannyGallivan]

I would expect the mid 80's Oilers, based mainly on scoring.

 

[Big Phil]

I would expect the mid 80's Oilers, based mainly on scoring.

They are in there, all of those great Oilers teams were on that list.

 

[666]

It might surprise you a bit. Or not. But either way, Tampa in 2019 is a surprising 35th on this list at +103. The 1977 Habs - no shock - lead this list with 216 and it isn't even close. Boston in 1971 is next with +192. Montreal in 1978 follows with +178. Actually, 5 of the top 6 teams are Habs teams.

Tampa, despite their huge year in 2019 isn't as high as you might think. Ottawa in 2006 is ahead of them, so is Chicago (no kidding) in 1974. As are the 1976 Islanders and the 1980 Sabres. Good teams, but still ones that surprise me that they are ahead of Tampa from this year.

Most of the top 40 are from the 1970s and 1980s based on the fact scoring was higher than it is today. Before Tampa in 2019 the highest was Detroit in 1996 with +144. Only the 1944, 1945 and 1959 Habs had at least +100 in goal differential pre-expansion. Two of them are war years, but 1959 is even more impressive since it was a 70 game schedule and teams played each other 14 times a year with lots of familiarity. The 1996 Wings are the only team in the 1990s with at least +100, which surprised me. The 2001 Devils are the only team in the 2000s after the 2006 Sens.

I am just thinking of other dominant teams that could have been here. The 1930 Bruins pro-rated definitely make this list but played just 44 games yet were +81. The teams that just miss the +100 cut are the 1993 Pens at +99. The 2006 Red Wings were +96. It surprised me that the 2002 Wings weren't even close.


Team Game Finder | Hockey-Reference.com

The 1977 Habs differential was 2.49 standard deviations above the mean.
The 2019 Lighting differential was 2.52 standard deviations above the mean so it was actually a bigger margin.

The Habs 387 goals were 2.57 standard deviations above the mean.
The Lighting's 325 goals were 2.65 standard deviations above the mean.

I believe that this is the best way to compare things across generations.

 

[MadLuke]

The 1977 Habs differential was 2.49 standard deviations above the mean.
The 2019 Lighting differential was 2.52 standard deviations above the mean so it was actually a bigger margin.

The Habs 387 goals were 2.57 standard deviations above the mean.
The Lighting's 325 goals were 2.65 standard deviations above the mean.

I believe that this is the best way to compare things across generations.

That seem a bit flawed, Lighting were #8 defensively, 30 goal worst than the number one team, while being #1 in scoring by 30 goals, i.e. flames offense with Islanders defence would have been has good.

While the Habs were the best offense by 64 goals and best defense by 22 goals, take the best non Habs team offense, give them the best non Habs team defense and that team would still be 86 goals worst that the Canadian that a year (a goal a game worst in differential).

I think a better way would be to remove the team when calculating standard deviation/average, after all +1000 team would create a giant standard deviation and raise the average a lot (so it would still be 0 like for every other case) and it's own dominance would hurt how much they appear dominant.

If we remove the Canadian's from the 1976-1977 season, goal differential std dev goes from 86 to 70 and the average is -12 instead of 0. giving their +216 performance a 3.26 instead 2.51 standard deviations above the mean.

Tampa bay is a bit strange when looking at the season result (2018-19 NHL Summary | Hockey-Reference.com) they +98 not +103, is there some goal during shootout not taking into account in one metric and in the other ?

Removing Tampa Bay, Std dev goes from 40.55 to 36.87 average to -3.266, making their +98 go from 2.41 standard deviation to +2.75, a good amount under the Habs domination.

 

[The Panther]

This list again shows that the early-80s' Canadiens were a much stronger team than is generally remembered. The 1980-81 squad is 39th all-time in goal differential, despite not having a single point-per-game scorer (except Lafleur, who played only 51 games). However, they were upset by Edmonton in the playoffs.

But the 1981-82 team was awesome. They finished just behind the Islanders and Edmonton at the top of the League, but had a better goal differential than either, and in fact were 10th best all time. Only 17 losses in 80 games. They had five players at +45 or better, the top three (Engblom, Langway, Robinson) being defenceman. Accordingly, the 1981-82 Habs were incredible defensively. Their League-best 223 goals against was as good or better than half a dozen teams this season, which is incredible for 1981-82 -- the highest scoring season in modern NHL history.

 

[VanIslander]

Let's also look at the opposite.

The 1938 Stanley Cup champion Blackhawks had the NHL's lowest goal total and second worst goals against for...

Minus-42 over a 48-game season.

Has any champion had a worse regular season goals differential?

 

[Albatros]

Interestingly Chicago also had a winning record only against the Maroons that season.

 

[BigBadBruins7708]

'71 Bruins...+192 goal differential, 2.46 goals per game better than their opponents

easily has to be the best to not win

 

[The Panther]

'71 Bruins...+192 goal differential, 2.46 goals per game better than their opponents

easily has to be the best to not win

Maybe, but I think in that era you have to look at the strength of competition. 4 of 14 teams that season were under .400. California was .288, and in six games against Boston they were outscored 33-8. I mean, Boston scored 399 goals that season, while California scored 199.

For that reason, I would tend to go with (in post-expansion era) Edmonton 1986, Pittsburgh 1993, or Detroit 1996 as the best teams not to win. But anyway, 1971 Bruins are certainly up there near the top.

 

[BigBadBruins7708]

Maybe, but I think in that era you have to look at the strength of competition. 4 of 14 teams that season were under .400. California was .288, and in six games against Boston they were outscored 33-8. I mean, Boston scored 399 goals that season, while California scored 199.

For that reason, I would tend to go with (in post-expansion era) Edmonton 1986, Pittsburgh 1993, or Detroit 1996 as the best teams not to win. But anyway, 1971 Bruins are certainly up there near the top.

True, but even with the lower competition they outpaced the field including the other "haves"

Bruins scored 399
2nd highest was 291 (Habs)

Bruins were +192
2nd highest was +93 (Hawks)

So while their raw numbers were inflated, they still ran away from everyone else scoring 24.5% more goals and having a differential 51.5% better than anyone else that year.

In scoring they were 2.10 standard deviations above the mean, and 2.27 standard deviations in +/-

 

[666]

That seem a bit flawed, Lighting were #8 defensively, 30 goal worst than the number one team, while being #1 in scoring by 30 goals, i.e. flames offense with Islanders defence would have been has good.

While the Habs were the best offense by 64 goals and best defense by 22 goals, take the best non Habs team offense, give them the best non Habs team defense and that team would still be 86 goals worst that the Canadian that a year (a goal a game worst in differential).

I think a better way would be to remove the team when calculating standard deviation/average, after a +1000 team would create a giant standard deviation and raise the average a lot and it's dominance would hurt how much they appear dominant.

If we remove the Canadian's from the 1976-1977 season, goal differential std dev goes from 86 to 70 and the average is -12 instead of 0. giving their +216 performance a 3.26 instead 2.51 standard deviations above the mean.

Tampa bay is a bit strange when looking at the season result (2018-19 NHL Summary | Hockey-Reference.com) they +98 not +103, is there some goal during shootout not taking into account in one metric and in the other ?

Removing Tampa Bay, Std dev goes from 40.55 to 36.87 average to -3.266, making their +98 go from 2.41 standard deviation to +2.75, a good amount under the Habs domination.

Yeah, I've thought about doing it this way too but eliminating the top performer from a normal distribution doesn't feel right. If you had scores in a classroom of 100, 60, 50, 40 would it be ok to eliminate the 100 and say that the guy who scored 40 was within one standard deviation. It's not my area of expertise but it doesn't feel right. I get where you're going however in that case I would remove the top performer and the lowest performer.

In that case I get 2.90 for Tampa(assuming 103) and 3.29 for the Habs.

 

[MadLuke]

Yeah, I've thought about doing it this way too but eliminating the top performer from a normal distribution doesn't feel right. If you had scores in a classroom of 100, 60, 50, 40 would it be ok to eliminate the 100 and say that the guy who scored 40 was within one standard deviation. It's not my area of expertise but it doesn't feel right. I get where you're going however in that case I would remove the top performer and the lowest performer.

In that case I get 2.90 for Tampa(assuming 103) and 3.29 for the Habs.

Yeah I am really unsure what you be the best here, my goal was not has much has removing outlier to have a clearer view, but removing just themselve because of much the best team ever would create by it's presence a lot of std dev, and everytime the performer metric is hurt by how good they are it does not feel right, imagine judging Gretzky ArtRoss dominance using the Top 10 scorer average/std dev and keeping is 200 points when everyone was under 120.

Specially the Habs were in a 18 teams league, their performance will affect the variable a lot.

 

[SealsFan]

I did actually go through all the seasons back to 1999 or so and count how many times a team reached 300 goals scored, but I forget how many, three maybe. So Tampa Bay is added to that very short list. I'll go back and count again when I get a chance.