HFBoards

Go Back   HFBoards > General Hockey Discussion > National Hockey League Talk
Mobile Hockey's Future Become a Sponsor Site Rules Support Forum vBookie Page 2
National Hockey League Talk Discuss NHL players, teams, games, and the Stanley Cup Playoffs.

Correlation between payroll and regualr season standing 2003/2004

Closed Thread
 
Thread Tools
Old
05-28-2004, 02:43 PM
  #1
CH
Registered User
 
Join Date: Jul 2003
Posts: 606
vCash: 500
Correlation between payroll and regular season standing 2003/2004

Just a little math exercize I went through to dispell a couple misconceptions about the NHL today.

I will explain what I did trying to keep it as simple as possible.

I ranked all 30 NHL teams in terms of payroll as of the beginning of the season (so as to not be too affected by trade deadline salary dumping) and I ranked all teams from 1 to 30 in terms of their final standing in the NHL this year (so playoffs will not be taken into account in this little study.

The rankings are as follows

Team Payroll Standings
Detroit 1 1
NY Rangers 2 25
Dallas 3 11
Philadelphia 4 8
Toronto 5 5
St Louis 6 14
Colorado 7 10
Anaheim 8 22
Washington 9 28
New Jersey 10 9
Los Angeles 11 20
Boston 12 4
NY Islanders 13 15
Montreal 14 13
Ottawa 15 6
Vancouver 16 7
Phoenix 17 26
Carolina 18 23
Calgary 19 12
San Jose 20 3
Tampa Bay 21 2
Buffalo 22 18
Columbus 23 27
Chicago 24 29
Edmonton 25 17
Atlanta 26 21
Pittsburgh 27 30
Florida 28 24
Minnesota 29 19
Nashville 30 16

Now I calculated the correlation between the payroll and standings number. This is a relatively simple statistics problem. The theory behind it would be covered in any first year university statistics course. A simple blurb about correlation can be found for example here

http://www.bized.ac.uk/timeweb/crunc...elate_expl.htm

Simply, if the two variables track exactly (ie highest payroll team has highest standings, 2nd highest payroll team 2nd in standings etc) we will get a value of 1 for the correlation.
If there is no tracking whatsoever between the variables we will get a correlation of zero.
If they anti-correlate (highest payroll last in standings, 2nd highest payroll 2nd last in standings etc) we will get a value of -1 for the correlation.

The correlation between payroll and standings in the 2003/2004 NHL season is 0.3717.

Now what does that tell us?

If we go back to the website I gave you to explain correlation and quote from it

Quote:

The strength and significance of the coefficient

The following general categories indicate a quick way of interpreting a calculated r value:

0.0 to 0.2 Very weak to negligible correlation
0.2 to 0.4 Weak, low correlation (not very significant)
0.4 to 0.7 Moderate correlation
0.7 to 0.9 Strong, high correlation
0.9 to 1.0 Very strong correlation
So in conclusion, there is a weak, low correlation between payroll and standings in the NHL last year. Its not very significant.

So why do so many posters on this board advocate major changes to the central bargaining agreement to get rid of a problem that is not very significant?

CH is offline  
Old
05-28-2004, 02:52 PM
  #2
Beukeboom Fan
Registered User
 
Beukeboom Fan's Avatar
 
Join Date: Feb 2002
Location: Chicago, IL
Posts: 11,724
vCash: 500
Quote:
Originally Posted by CH
Just a little math exercize I went through to dispell a couple misconceptions about the NHL today.

I will explain what I did trying to keep it as simple as possible.

I ranked all 30 NHL teams in terms of payroll as of the beginning of the season (so as to not be too affected by trade deadline salary dumping) and I ranked all teams from 1 to 30 in terms of their final standing in the NHL this year (so playoffs will not be taken into account in this little study.

The rankings are as follows

Team Payroll Standings
Detroit 1 1
NY Rangers 2 25
Dallas 3 11
Philadelphia 4 8
Toronto 5 5
St Louis 6 14
Colorado 7 10
Anaheim 8 22
Washington 9 28
New Jersey 10 9
Los Angeles 11 20
Boston 12 4
NY Islanders 13 15
Montreal 14 13
Ottawa 15 6
Vancouver 16 7
Phoenix 17 26
Carolina 18 23
Calgary 19 12
San Jose 20 3
Tampa Bay 21 2
Buffalo 22 18
Columbus 23 27
Chicago 24 29
Edmonton 25 17
Atlanta 26 21
Pittsburgh 27 30
Florida 28 24
Minnesota 29 19
Nashville 30 16

Now I calculated the correlation between the payroll and standings number. This is a relatively simple statistics problem. The theory behind it would be covered in any first year university statistics course. A simple blurb about correlation can be found for example here

http://www.bized.ac.uk/timeweb/crunc...elate_expl.htm

Simply, if the two variables track exactly (ie highest payroll team has highest standings, 2nd highest payroll team 2nd in standings etc) we will get a value of 1 for the correlation.
If there is no tracking whatsoever between the variables we will get a correlation of zero.
If they anti-correlate (highest payroll last in standings, 2nd highest payroll 2nd last in standings etc) we will get a value of -1 for the correlation.

The correlation between payroll and standings in the 2003/2004 NHL season is 0.3717.

Now what does that tell us?

If we go back to the website I gave you to explain correlation and quote from it



So in conclusion, there is a weak, low correlation between payroll and standings in the NHL last year. Its not very significant.

So why do so many posters on this board advocate major changes to the central bargaining agreement to get rid of a problem that is not very significant?
Hate to volunteer you for more work, but how about expanding the sample size? Say over the last 5-10 years, what was the correlation between player salaries and winning the Cup? Or player salaries and making the play-off's?

Beukeboom Fan is offline  
Old
05-28-2004, 02:52 PM
  #3
djhn579
Registered User
 
Join Date: Mar 2003
Location: Tonawanda, NY
Country: United States
Posts: 1,747
vCash: 500
Quote:
Originally Posted by CH
Just a little math exercize I went through to dispell a couple misconceptions about the NHL today.

I will explain what I did trying to keep it as simple as possible.

I ranked all 30 NHL teams in terms of payroll as of the beginning of the season (so as to not be too affected by trade deadline salary dumping) and I ranked all teams from 1 to 30 in terms of their final standing in the NHL this year (so playoffs will not be taken into account in this little study.

The rankings are as follows

Team Payroll Standings
Detroit 1 1
NY Rangers 2 25
Dallas 3 11
Philadelphia 4 8
Toronto 5 5
St Louis 6 14
Colorado 7 10
Anaheim 8 22
Washington 9 28
New Jersey 10 9
Los Angeles 11 20
Boston 12 4
NY Islanders 13 15
Montreal 14 13
Ottawa 15 6
Vancouver 16 7
Phoenix 17 26
Carolina 18 23
Calgary 19 12
San Jose 20 3
Tampa Bay 21 2
Buffalo 22 18
Columbus 23 27
Chicago 24 29
Edmonton 25 17
Atlanta 26 21
Pittsburgh 27 30
Florida 28 24
Minnesota 29 19
Nashville 30 16

Now I calculated the correlation between the payroll and standings number. This is a relatively simple statistics problem. The theory behind it would be covered in any first year university statistics course. A simple blurb about correlation can be found for example here

http://www.bized.ac.uk/timeweb/crunc...elate_expl.htm

Simply, if the two variables track exactly (ie highest payroll team has highest standings, 2nd highest payroll team 2nd in standings etc) we will get a value of 1 for the correlation.
If there is no tracking whatsoever between the variables we will get a correlation of zero.
If they anti-correlate (highest payroll last in standings, 2nd highest payroll 2nd last in standings etc) we will get a value of -1 for the correlation.

The correlation between payroll and standings in the 2003/2004 NHL season is 0.3717.

Now what does that tell us?

If we go back to the website I gave you to explain correlation and quote from it



So in conclusion, there is a weak, low correlation between payroll and standings in the NHL last year. Its not very significant.

So why do so many posters on this board advocate major changes to the central bargaining agreement to get rid of a problem that is not very significant?

Just like +/-, some stats have to be looked at over the long term to be meaningful. Try doing that over a 5 or 10 year period and see what you get. I'm sure that you will find that it becomes more meaningful. Unfortunately, I'm not a statistics major, so don't ask me to do it myself...

And the central argument is that certain teams drive up salaries for everybody else, making it more difficult to remain competetive or if you don't like that, hold on to their best players which goes towards maintaining your fan base.

djhn579 is offline  
Old
05-28-2004, 02:59 PM
  #4
littleHossa
Registered User
 
Join Date: Apr 2003
Location: Ottawa
Country: Canada
Posts: 1,753
vCash: 500
Interesting to see that after Nashville, Tampa Bay is the playoff team with the smallest payroll.

littleHossa is offline  
Old
05-28-2004, 03:02 PM
  #5
JCD
Registered User
 
JCD's Avatar
 
Join Date: Feb 2002
Country:
Posts: 14,432
vCash: 500
As noted, using a sample size of a single year carries very little significance.

You need to look at a 5-10 year sample to draw any meaningful conclusion.

JCD is offline  
Old
05-28-2004, 03:06 PM
  #6
degroat*
 
Join Date: Jun 2002
Location: http://nhl.degroat.n
Posts: 8,108
vCash: 500
I'm sorry you went through all that work, but your findings aren't very significant because they actual correlation is between payroll and sustained success.

Any one team can have a good season regardless of the payroll while a lot of things would have to fall into place for any team to be successful over a few years with a low payroll.

degroat* is offline  
Old
05-28-2004, 03:08 PM
  #7
CH
Registered User
 
Join Date: Jul 2003
Posts: 606
vCash: 500
Quote:
Originally Posted by JCD
As noted, using a sample size of a single year carries very little significance.

You need to look at a 5-10 year sample to draw any meaningful conclusion.
I don't have any pre-season payroll numbers for any years other than this one.

Can anybody point me toward some?



Further, I think that this is a meaningful result. In a 30 team league, there are 29 degrees of freedom which gives OK statistics.

I do understand that if you disagree with the result, you must find something to complain about instead of considering the fact that your preconceived ideas may be incorrect.

CH is offline  
Old
05-28-2004, 03:11 PM
  #8
YKOil
Registered User
 
Join Date: Apr 2004
Posts: 1,232
vCash: 500
In addition to using several years for your sample, try eliminating the two worst outliers (at either end of the spectrum) in your yearly sample.

That would probably include the Rangers and Ottawa every year.


YKOil

YKOil is offline  
Old
05-28-2004, 03:11 PM
  #9
JCD
Registered User
 
JCD's Avatar
 
Join Date: Feb 2002
Country:
Posts: 14,432
vCash: 500
Quote:
Originally Posted by CH
I don't have any pre-season payroll numbers for any years other than this one.

Can anybody point me toward some?

Further, I think that this is a meaningful result. In a 30 team league, there are 29 degrees of freedom which gives OK statistics.

I do understand that if you disagree with the result, you must find something to complain about instead of considering the fact that your preconceived ideas may be incorrect.
Actually, this is not statistically valid to show a trend. For this one year, it holds up, but one year alone. But if we are going to talk about trend over time, you need more than a 1-year snapshot. A one-year abberation can completley skew this analysis, those get washed out over time.

Get to 5 years and I guarentee you see a strong corrolation between payroll and points.

I have a bunch of old 'Bucks and Pucks' tucked away in a box somewhere, but am in the middle of moving so have no idea where they are. I am sure the data is on-line somewhere.

JCD is offline  
Old
05-28-2004, 03:18 PM
  #10
JCD
Registered User
 
JCD's Avatar
 
Join Date: Feb 2002
Country:
Posts: 14,432
vCash: 500
http://www.hockeyzoneplus.com/$maseq_e.htm

I don't know the website, so cannot vouch for it's reliability. That said, I have no reason to believe it is false.

JCD is offline  
Old
05-28-2004, 03:23 PM
  #11
Beukeboom Fan
Registered User
 
Beukeboom Fan's Avatar
 
Join Date: Feb 2002
Location: Chicago, IL
Posts: 11,724
vCash: 500
Quote:
Originally Posted by CH
I don't have any pre-season payroll numbers for any years other than this one.

Can anybody point me toward some?



Further, I think that this is a meaningful result. In a 30 team league, there are 29 degrees of freedom which gives OK statistics.

I do understand that if you disagree with the result, you must find something to complain about instead of considering the fact that your preconceived ideas may be incorrect.
I don't know of where you can find previous season's payroll totals.

But instead of being a smart-ass (last parpagraph), how about realizing that you have a very small sample size, and that many other factors impact the results? Things such as expansion, injuries, & coaching all play a major impact on the standing, and skew any single season results.

Just off the top of my head - what are the 4 teams that have won the last 10 Stanley Cups? DET-3, COL- 2, DAL- 1, NJ- 3, NYR-1. I would bet that of those 10 teams, the only one that might of been below the median was the 1st NJ Cup winner, but I wouldn't place a bet on that.

Which teams have been the most consistent over the last 10 years? How about COL with 10 straight division championships? Just refresh my memory - what's their payroll look like? How about PHI? How about DAL (with a minor blip or two). Detroit? Toronto? Or on the other end - how did NASH/CLM/OTT/TB all fair after expansion when they had a low payroll?

There are obvious exceptions - the recent Rangers, or teams like MINN whose coaching delivered a lot better results than the payroll deserved.

Beukeboom Fan is offline  
Old
05-28-2004, 03:24 PM
  #12
YellHockey*
 
Join Date: Feb 2002
Posts: 2,830
vCash: 500
Quote:
Originally Posted by JCD
Get to 5 years and I guarentee you see a strong corrolation between payroll and points.
You should see a stronger correlation because of the effect of expansion teams.

Expansion teams are typically comprised of youngsters and journeymen, neither of which makes much money. And those teams also don't typically win many games.

YellHockey* is offline  
Old
05-28-2004, 03:26 PM
  #13
YellHockey*
 
Join Date: Feb 2002
Posts: 2,830
vCash: 500
Quote:
Originally Posted by Beukeboom Fan
But instead of being a smart-ass (last parpagraph), how about realizing that you have a very small sample size, and that many other factors impact the results? Things such as expansion, injuries, & coaching all play a major impact on the standing, and skew any single season results.

Just off the top of my head - what are the 4 teams that have won the last 10 Stanley Cups? DET-3, COL- 2, DAL- 1, NJ- 3, NYR-1. I would bet that of those 10 teams, the only one that might of been below the median was the 1st NJ Cup winner, but I wouldn't place a bet on that.
You're bashing him for his sample size but you're looking at the playoffs (which have a much smaller sample size then the regular season) to validate your opinions?

YellHockey* is offline  
Old
05-28-2004, 03:36 PM
  #14
Beukeboom Fan
Registered User
 
Beukeboom Fan's Avatar
 
Join Date: Feb 2002
Location: Chicago, IL
Posts: 11,724
vCash: 500
Quote:
Originally Posted by BlackRedGold
You're bashing him for his sample size but you're looking at the playoffs (which have a much smaller sample size then the regular season) to validate your opinions?
I'm looking at 10 years of data (based on my memory, so it's not validated by any stretch of the imagination) to develop a trend. He looked at one year, and tried to infer that it related to the entire population (all years).

My point was that the teams that can afford to pay the most, have the most success. By success, I mean winning in the play-off's. You occasionally have teams that catch lightning in a bottle (TB, CAL, ANA, CAR, FLA), but they are relatively rare, and can ussually point to one of two factors for their success - goaltending or coaching.

Every team that has consistently produced over the last 10 years is in the upper 1/2 to 1/3 of team payrolls. DET, COL, NJ, PHI, DAL. The 2 young teams that you might be able to mention going forward are TB & OTT, and I think that both of their payroll's are going way up if the core of the team is kept together.

Beukeboom Fan is offline  
Old
05-28-2004, 03:45 PM
  #15
Tom_Benjamin
Registered User
 
Join Date: Sep 2003
Posts: 1,152
vCash: 500
Rod Bryden made this case many times during the Senator saga. He kept saying payroll correlated very weakly with success. (It is the only thing he ever said that I agreed with.) Expanding the sample size won't change anything. There are always big payroll teams that tank and there are always small payroll teams that surprise.

Another way to look at the data is to compare the change in payroll to the change in the standings. The correlation was actually negative this season. Eleven teams cut payroll (by a total of about $60 million) last season and collectively gained 28 points in the standings. The other 19 teams increased payroll (by a total of about $105 million) and collectively lost 50 points in the standings.

Tom

Tom_Benjamin is offline  
Old
05-28-2004, 03:53 PM
  #16
Karl Pilkington
Registered User
 
Karl Pilkington's Avatar
 
Join Date: Feb 2004
Location: Ottawa
Country: Canada
Posts: 2,990
vCash: 500
Quote:
Originally Posted by CH
Just a little math exercize I went through to dispell a couple misconceptions about the NHL today.

I will explain what I did trying to keep it as simple as possible.

I ranked all 30 NHL teams in terms of payroll as of the beginning of the season (so as to not be too affected by trade deadline salary dumping) and I ranked all teams from 1 to 30 in terms of their final standing in the NHL this year (so playoffs will not be taken into account in this little study.

The rankings are as follows

Team Payroll Standings
Detroit 1 1
NY Rangers 2 25
Dallas 3 11
Philadelphia 4 8
Toronto 5 5
St Louis 6 14
Colorado 7 10
Anaheim 8 22
Washington 9 28
New Jersey 10 9
Los Angeles 11 20
Boston 12 4
NY Islanders 13 15
Montreal 14 13
Ottawa 15 6
Vancouver 16 7
Phoenix 17 26
Carolina 18 23
Calgary 19 12
San Jose 20 3
Tampa Bay 21 2
Buffalo 22 18
Columbus 23 27
Chicago 24 29
Edmonton 25 17
Atlanta 26 21
Pittsburgh 27 30
Florida 28 24
Minnesota 29 19
Nashville 30 16

Now I calculated the correlation between the payroll and standings number. This is a relatively simple statistics problem. The theory behind it would be covered in any first year university statistics course. A simple blurb about correlation can be found for example here

http://www.bized.ac.uk/timeweb/crunc...elate_expl.htm

Simply, if the two variables track exactly (ie highest payroll team has highest standings, 2nd highest payroll team 2nd in standings etc) we will get a value of 1 for the correlation.
If there is no tracking whatsoever between the variables we will get a correlation of zero.
If they anti-correlate (highest payroll last in standings, 2nd highest payroll 2nd last in standings etc) we will get a value of -1 for the correlation.

The correlation between payroll and standings in the 2003/2004 NHL season is 0.3717.

Now what does that tell us?

If we go back to the website I gave you to explain correlation and quote from it



So in conclusion, there is a weak, low correlation between payroll and standings in the NHL last year. Its not very significant.

So why do so many posters on this board advocate major changes to the central bargaining agreement to get rid of a problem that is not very significant?


I did something very similar for my grade 12 Data Management project I'm working on... I came up with an equation to set a fair payroll but explained that the only benefits from this would be that the owners would save money. The league would remain as competetive as it is now.

Your explanation using correlation however is excellent... can I borrow some of your analysis for my project?

Karl Pilkington is offline  
Old
05-28-2004, 03:53 PM
  #17
YellHockey*
 
Join Date: Feb 2002
Posts: 2,830
vCash: 500
Quote:
Originally Posted by Beukeboom Fan
I'm looking at 10 years of data (based on my memory, so it's not validated by any stretch of the imagination) to develop a trend. He looked at one year, and tried to infer that it related to the entire population (all years).
If you're going by just the playoffs for 10 years, that is at most 1050 games.

(8 first rounds + 4 second rounds + 2 third rounds + 1 final) * 7 games * 10 years

While he is going by 1215 games.

82 games * 30 teams / 2 (since there are two teams in each game)

I'd say his results are more statistically significant.

YellHockey* is offline  
Old
05-28-2004, 04:07 PM
  #18
CH
Registered User
 
Join Date: Jul 2003
Posts: 606
vCash: 500
Slowly I am working on giving you correlations between payroll and regular season standings for more than just one season. I have now calculated the 2002/03 season.

Team Payroll Standing
NY Rangers 1 19
Detroit 2 3
St Louis 3 8
Dallas 4 2
Colorado 5 6
Philadelphia 6 5
Toronto 7 9
New Jersey 8 4
Washington 9 13
Montreal 10 21
San Jose 11 25
Chicago 12 17
Phoenix 13 20
Los Angeles 14 18
NY Islanders 15 16
Carolina 16 30
Anaheim 17 11
Boston 18 15
Calgary 19 22
Florida 20 27
Vancouver 21 7
Pittsburgh 22 29
Buffalo 23 26
Edmonton 24 14
Ottawa 25 1
Tampa Bay 26 12
Columbus 27 28
Atlanta 28 23
Nashville 29 24
Minnesota 30 10

Correlation this season = 0.4047 which is a bit higher. It barely makes tyhe moderate correlation category.

Correlation in the two seasons put together = 0.3816.

As we get closer to expansion I would expect to see moderate increases in the correlation between payroll and standings because expansion teams will all have very low payroll and very low positions in the standings.

When I get time I will compute this for a few more seasons going back in time. Likely it wont be immediately due to the approaching long weekend in the US (if time permits I may get one more done tonight).

Still the correlation between payroll and success in the standings in the NHL is pretty low over the last two seasons.

CH is offline  
Old
05-28-2004, 04:19 PM
  #19
Beukeboom Fan
Registered User
 
Beukeboom Fan's Avatar
 
Join Date: Feb 2002
Location: Chicago, IL
Posts: 11,724
vCash: 500
Quote:
Originally Posted by BlackRedGold
If you're going by just the playoffs for 10 years, that is at most 1050 games.

(8 first rounds + 4 second rounds + 2 third rounds + 1 final) * 7 games * 10 years

While he is going by 1215 games.

82 games * 30 teams / 2 (since there are two teams in each game)

I'd say his results are more statistically significant.
I'll say it again - you cannot make a determination between payroll & standings based on one years results. I don't care how many individual games took place in this season, it does not give you enough data to say that that there is no correlation between payroll and results.

Beukeboom Fan is offline  
Old
05-28-2004, 04:20 PM
  #20
degroat*
 
Join Date: Jun 2002
Location: http://nhl.degroat.n
Posts: 8,108
vCash: 500
Once again.... the correlation is between payroll and sustained success.

Any stats you put together for any individual season are meaningless. Take the team payrolls from the last 5 seasons, add them together. Do the same for the point totals. Then, ta-dow!

degroat* is offline  
Old
05-28-2004, 04:23 PM
  #21
YellHockey*
 
Join Date: Feb 2002
Posts: 2,830
vCash: 500
Quote:
Originally Posted by Stich
Once again.... the correlation is between payroll and sustained success.

Any stats you put together for any individual season are meaningless.
And how do you define sustained success?

The correlation for an individual season is not meaningless. If the NHL had competitive balance problems as bad as the owners make it out to be, there would be a strong correlation for every season.

YellHockey* is offline  
Old
05-28-2004, 04:24 PM
  #22
YellHockey*
 
Join Date: Feb 2002
Posts: 2,830
vCash: 500
Quote:
Originally Posted by Beukeboom Fan
I don't care how many individual games took place in this season, it does not give you enough data to say that that there is no correlation between payroll and results.
If that's the case, prove it!

How much data is enough? How big of a sample size is statistically significant for you?

If you don't know how to figure it out, then you aren't in a position to judge.

YellHockey* is offline  
Old
05-28-2004, 04:34 PM
  #23
degroat*
 
Join Date: Jun 2002
Location: http://nhl.degroat.n
Posts: 8,108
vCash: 500
Quote:
Originally Posted by BlackRedGold
And how do you define sustained success?
I just told him exactly what to do... can't make it any simpler.

Quote:
Originally Posted by BlackRedGold
The correlation for an individual season is not meaningless. If the NHL had competitive balance problems as bad as the owners make it out to be, there would be a strong correlation for every season.
The competitive balance problems are as bad as the owners make it out to be. There are two different situations that can result in a low payroll team being successful.

One of those two has already been pointed out in this thread and that's with a great coach or a great goaltender. Teams such as Carolina, Minnesota, and Anaheim are examples of this type which usually only lasts for one season.

The other situation is flawless management... the nearly-perfect signings, nearly-prefect trades, nearly-prefect drafting. This is what you see happening in Tampa and Ottawa. This is more rare and when it a team does accomplish this they're usually able to sustain it for a few years... until their star players hit their payday. If all the Lightning players were free agents this summer, they'd never be able to keep the team together.

The thing is though... it's not completely about competitive balance. It's about how the fans preceive the chances are of their favorite team.

Let's say that you're an (keyword...) average fan of the Florida Panthers... would you want to buy season tickets for a team that may or... more likely... may not make the playoffs in any given year? Not likely.

As far as owners are concerned, it's better for ticket sales for the fans to preceive the team as a good team than it is for the team to actually be good. This is why a team like San Jose averaged only ~15K per game this past year and is the reason why the Rangers never have a problem filling the arena.

In other words... a system that gives fans hope is what's best for the league financially.

degroat* is offline  
Old
05-28-2004, 04:35 PM
  #24
Beukeboom Fan
Registered User
 
Beukeboom Fan's Avatar
 
Join Date: Feb 2002
Location: Chicago, IL
Posts: 11,724
vCash: 500
Quote:
Originally Posted by BlackRedGold
If that's the case, prove it!

How much data is enough? How big of a sample size is statistically significant for you?

If you don't know how to figure it out, then you aren't in a position to judge.
I know enough statistics to be dangerous and nothing else. I can throw out Z-scores and the like, but I don't have the time to do an in depth statistical study.

I do know enough about statistic though to say that you CANNOT prove a correlation (or lack there-of) based on ONE or TWO years of data. It doesn't matter that there are 1000 or 1000000 games played in those years, you are looking at ONE data set.

I do know that the sample size is based on the population, and we are looking at an incredibly small sample (really the last 15 or so years is when financial matters really started to impact the NHL).

Especially when there are known cases that skew the data. Examples of this are injuries (Example - Kings), coaching (MINN), and stupidity (Rangers). In fact, look at the data and you can identify the 3 or 4 major cases that are skewing the data (OTT, MINN, NYR & TB).

Just look at the data though - 8 of the 10 top teams made the play-off's.

Beukeboom Fan is offline  
Old
05-28-2004, 05:12 PM
  #25
CH
Registered User
 
Join Date: Jul 2003
Posts: 606
vCash: 500
Third year (and final ... at least for now) 2001/02 season.

Team Payroll Standing
Detroit 1 1
St Louis 2 6
NY Rangers 3 21
Dallas 4 18
Philadelphia 5 7
Washington 6 19
Colorado 7 4
Toronto 8 3
San Jose 9 5
Los Angeles 10 12
Boston 11 2
New Jersey 12 10
Florida 13 28
Chicago 14 9
Montreal 15 17
Anaheim 16 24
NY Islanders 17 8
Carolina 18 16
Pittsburgh 19 26
Phoenix 20 11
Vancouver 21 13
Buffalo 22 20
Calgary 23 22
Ottawa 24 14
Tampa Bay 25 27
Edmonton 26 15
Columbus 27 29
Atlanta 28 30
Nashville 29 25
Minnesota 30 23

Correlation this season =0.6111 This is a moderate correlation. Clearly as we get closer to expansion, correlation is continuing to go up - and at a faster rate. All of the expansion teams have the lowest payrolls and also the worst records. So expansion creates a competitve balance problem. Want to solve that? Keep the same CBA but do not expand the league.

To summarize

2003/04 correlation between payroll and standings = 0.3717
2002/03 correlation = 0.4047
2001/02 correlation = 0.6111

total for the three seasons 0.4521

I think expansion created any competive balance problems in the past. That is definitely consistent with the data. These problems were moderate but and becoming less and less significant as time passes from expansion.

This might be what some fans are pointing at when they complain about competeive balance problems (or they may just be complaining). It is important for them to notice that this problem is fixing itself with no CBA change.

CH is offline  
Closed Thread

Forum Jump


Bookmarks

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off



All times are GMT -5. The time now is 09:36 AM.

monitoring_string = "e4251c93e2ba248d29da988d93bf5144"
Contact Us - HFBoards - Archive - Privacy Statement - Terms of Use - Advertise - Top - AdChoices

vBulletin Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
HFBoards.com is a property of CraveOnline Media, LLC, an Evolve Media, LLC company. 2014 All Rights Reserved.