HFBoards Best Team to not win a Cup
 Register FAQ Members List Search Today's Posts Mark Forums Read
 Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie Page 2
 Notices Please do not post or solicit links to illegal game streams.

 The History of Hockey Relive great moments in hockey history and discuss how the game has changed over time.

# Best Team to not win a Cup

06-08-2005, 11:33 PM
#76
Ogopogo*

Join Date: Apr 2005
Location: Edmonton
Country:
Posts: 15,951
vCash: 500
Quote:
 Originally Posted by moneyp I'll use a six-team league to save myself some typing. The NHL in 1943-44: Team - GF - GA - Dif/G Boston - 223 - 268 - -0.90 Chicago - 178 - 187 - -0.18 Detroit - 214 - 177 - +0.74 Montreal - 234 - 109 - +2.50 NY Rangers - 162 - 310 - -2.96 Toronto - 214 - 174 - +0.80 You find the standard deviation for all of the teams goal differentials. Standard deviation explains how spread out the results of any group of data is (further explanation can be found here). I won't print the calculation, which you'd be crazy to figure out manually any way. It's the STDEVP function in Excel. The result for 1943-44 is 1.69, which is the second-highest result ever (behind 1920). Montreal was great in 1943-44, but the Rangers were actually as bad (even worse) than Montreal was good. Playing one-fifth of your games against those Rangers will make any team look better than it is. So the question becomes, how difficult was the achievement of having a +2.5 goal differential given the way the rest of the league performed that year? That's what standard deviation scores answer. I explained this on the other thread, but one more time: The math for figuring out standard deviation scores is (A minus B) divided by C. A = the team's goal differential B = the mean (or average) of the league's goal differential. Because every goal scored means a goal against for another team, the result for this should always be zero. C = The standard deviation of goal differentials (as discussed above) for the entire league that year. Thankfully, Excel has a STANDARDIZE function for this as well. To simplify, it takes the team achievement (goal differential, in this case) and compares it to the spread of that achievement across the league, bringing the extremes of the data into focus. The SDS scores for 1943-44: Boston: -0.53 Chicago: -0.11 Detroit: +0.44 Montreal: +1.48 NY Rangers: -1.76 Toronto: +0.47 Montreal's total ranks 70th all-time. They were actually better in 1945, with a +2.14 goal differential and a 1.56 SDS (54th all-time). The 1944 NY Rangers are the 45th worst team of all-time by these rankings. Maybe I'll post that list at some point.
Thank you.

 06-09-2005, 07:29 AM #77 mcphee Registered User     Join Date: Feb 2003 Posts: 19,101 vCash: 500 Damn math guys,making me feel dumb[er]. I've got a few non standard deviations myself, hence my autographed collection of restraining orders.
06-09-2005, 11:31 AM
#78
Snap Wilson
Registered User

Join Date: Sep 2003
Posts: 5,838
vCash: 500
Quote:
 Originally Posted by gary69 Interestingly, but perhaps not surprisingly, the then Rangers starting goalie Ken McAuley has the worst goalie stats ever in NHL history for a starting goalie, which he was for two seasons, with a decent number of games (96 games, GAA 5.61).
My old man used to say that the only player on the Rangers that year that could have played for the Habs was Hextall.

Quote:
 Originally Posted by gary69 Maybe you've posted them before, but I've missed them, anyway it would be nice to see the league-wide standard deviations for every year. I take one could draw conclusions from this, how even the league was during each period. Please mention also, if the reason for high standard devision in due to only one (or very small portion of teams).
It would never be due to just one team. If five of six teams (for example) were closely grouped, and one was really good or really lousy, you'd have a low standard deviation.

Here are the standard deviations for the NHL 1918 to present. Remember, the higher the number, the more unbalanced the league.

1918: 0.83
1919: 1.07
1920: 2.19 (welcome to the league, Quebec Bulldogs)
1921: 0.99
1922: 0.59
1923: 0.82
1924: 0.73
1925: 1.21 (welcome to the league, Maroons and Bruins)
1926: 0.62
1927: 0.45
1928: 0.86
1929: 0.58
1930: 0.97
1931: 1.02
1932: 0.48
1933: 0.54
1934: 0.50
1935: 0.72
1936: 0.39
1937: 0.45
1938: 0.79
1939: 0.89
1940: 1.05
1941: 0.99
1942: 0.65
1943: 0.91
1944: 1.69

I remember looking into the great jump here, but I don't remember what my conclusions were. Teams losing players to wartime, I think. This was also the year that the redline started and the Habs had a team poised to take advantage of it.

1945: 1.37
1946: 0.55
1947: 0.72
1948: 0.48
1949: 0.55
1950: 0.51
1951: 0.98
1952: 0.73
1953: 0.64
1954: 0.81
1955: 0.81
1956: 0.73
1957: 0.60
1958: 0.64
1959: 0.67
1960: 0.59
1961: 0.75
1962: 1.02 (an interesting, but apparently completely random blip)
1963: 0.61
1964: 0.57
1965: 0.79
1966: 0.88
1967: 0.71
1968: 0.50

Now this is interesting. The NHL's great expansion year, standard deviation actually went down to a historically low level. In part, this is because the league actually doubled in size, meaning that all of the expansion teams got to play other expansion teams quite a bit, making their records look pretty decent. But in any case, the existing teams didn't roll over the newcomers, and only one team (the Oakland Seals) had a notably lousy record.

1969: 0.73
1970: 0.81
1971: 1.08

And we enter the expansion era. Fourteen teams...

1972: 1.00
1973: 1.03

Sixteen teams...

1974: 0.95
1975: 1.30

Eighteen teams...

1976: 1.16
1977: 1.05
1978: 1.09
1979: 0.87

Back to seventeen...

1980: 0.70

And up to twenty-one. Again, we see a drop in standard deviation in an expansion year, but this was offset by the best of the WHA coming into the league.

1981: 0.82
1983: 0.91
1984: 0.87
1985: 0.84
1986: 0.80
1987: 0.44

Lowest SD ever for the league. A great year, very competitive. I was traveling the country at that point, and managed to hit fifteen of the twenty-one arenas that year.

1988: 0.62
1989: 0.64
1990: 0.62
1991: 0.67
1992: 0.60
1993: 0.92
1994: 0.69

Five teams were added over the past three seasons but only '93 shows any effect. The advancements in defensive play (or prominence of clutch-and-grab play, if you're a cynic) have the effect of neutralising the more offensively-talented teams in the league, and increasing parity.

1995: 0.65
1996: 0.70
1997: 0.46
1998: 0.53
1999: 0.52
2000: 0.60
2001: 0.59
2002: 0.51
2003: 0.52
2004: 0.55

And the league has more or less remained static since then. Should we see rules that actually increase offensive freedom, I predict a rise in standard deviation. We'll see.

06-09-2005, 04:21 PM
#79
gary69
Registered User

Join Date: Sep 2004
Location: Then and there
Posts: 3,182
vCash: 500
Quote:
 Originally Posted by moneyp My old man used to say that the only player on the Rangers that year that could have played for the Habs was Hextall. It would never be due to just one team. If five of six teams (for example) were closely grouped, and one was really good or really lousy, you'd have a low standard deviation. Here are the standard deviations for the NHL 1918 to present. Remember, the higher the number, the more unbalanced the league. 1918: 0.83 1919: 1.07 1920: 2.19 (welcome to the league, Quebec Bulldogs) 1921: 0.99 1922: 0.59 1923: 0.82 1924: 0.73 1925: 1.21 (welcome to the league, Maroons and Bruins) 1926: 0.62 1927: 0.45 1928: 0.86 1929: 0.58 1930: 0.97 1931: 1.02 1932: 0.48 1933: 0.54 1934: 0.50 1935: 0.72 1936: 0.39 1937: 0.45 1938: 0.79 1939: 0.89 1940: 1.05 1941: 0.99 1942: 0.65 1943: 0.91 1944: 1.69 I remember looking into the great jump here, but I don't remember what my conclusions were. Teams losing players to wartime, I think. This was also the year that the redline started and the Habs had a team poised to take advantage of it. 1945: 1.37 1946: 0.55 1947: 0.72 1948: 0.48 1949: 0.55 1950: 0.51 1951: 0.98 1952: 0.73 1953: 0.64 1954: 0.81 1955: 0.81 1956: 0.73 1957: 0.60 1958: 0.64 1959: 0.67 1960: 0.59 1961: 0.75 1962: 1.02 (an interesting, but apparently completely random blip) 1963: 0.61 1964: 0.57 1965: 0.79 1966: 0.88 1967: 0.71 1968: 0.50 Now this is interesting. The NHL's great expansion year, standard deviation actually went down to a historically low level. In part, this is because the league actually doubled in size, meaning that all of the expansion teams got to play other expansion teams quite a bit, making their records look pretty decent. But in any case, the existing teams didn't roll over the newcomers, and only one team (the Oakland Seals) had a notably lousy record. 1969: 0.73 1970: 0.81 1971: 1.08 And we enter the expansion era. Fourteen teams... 1972: 1.00 1973: 1.03 Sixteen teams... 1974: 0.95 1975: 1.30 Eighteen teams... 1976: 1.16 1977: 1.05 1978: 1.09 1979: 0.87 Back to seventeen... 1980: 0.70 And up to twenty-one. Again, we see a drop in standard deviation in an expansion year, but this was offset by the best of the WHA coming into the league. 1981: 0.82 1983: 0.91 1984: 0.87 1985: 0.84 1986: 0.80 1987: 0.44 Lowest SD ever for the league. A great year, very competitive. I was traveling the country at that point, and managed to hit fifteen of the twenty-one arenas that year. 1988: 0.62 1989: 0.64 1990: 0.62 1991: 0.67 1992: 0.60 1993: 0.92 1994: 0.69 Five teams were added over the past three seasons but only '93 shows any effect. The advancements in defensive play (or prominence of clutch-and-grab play, if you're a cynic) have the effect of neutralising the more offensively-talented teams in the league, and increasing parity. 1995: 0.65 1996: 0.70 1997: 0.46 1998: 0.53 1999: 0.52 2000: 0.60 2001: 0.59 2002: 0.51 2003: 0.52 2004: 0.55 And the league has more or less remained static since then. Should we see rules that actually increase offensive freedom, I predict a rise in standard deviation. We'll see.
Thanks, I would have thought the expansions to have had a bigger impact on league balance. League seems to have been a pretty well balanced for decades now, now everybody can get back to arguing why has/is it been so...

I tend to agree with your notes on clutching and grabbing. Since the 80's offensive league seems to have been reasonably well balanced (even if not quite what it is now), hopefully they could do it again with increased offense.

Last edited by gary69: 06-12-2005 at 02:59 PM.

Forum Jump