Explain this present Shootout anomaly.. East vs West
- Thread starter DL44
- Start date
CanadienShark
Registered User
- Dec 18, 2012
- 37,554
- 10,826
I may be way off base, but the West is more defensive - like 1-0, 2-1 games rather than the higher scoring East - with 5-3, 4-2 type games.
DL44
Status quo
If i would take a shot at it..
The standings in the West is much less likely to be representative of the final standings than the East presently based on the strength of the ROWs.
With a much larger SO effect going on in the West, it will take a while longer for the cream to rise as the SO effect decreases over the course of the yr.
As to Why its happening now... maybe an effect of a more balanced conference?
Or are there personnel issues creating this unique situation?
That's fair. But if it ends up balancing as the season goes on, then that can be waved off.
Thinking along the same lines - more explosive players in the East coming thru and scoring to finish games thru weaker D teams?
Again may prove false if numbers end up balancing out... but presently 3 vs 15 is huge difference.
The standings in the West is much less likely to be representative of the final standings than the East presently based on the strength of the ROWs.
With a much larger SO effect going on in the West, it will take a while longer for the cream to rise as the SO effect decreases over the course of the yr.
As to Why its happening now... maybe an effect of a more balanced conference?
Or are there personnel issues creating this unique situation?
That's fair. But if it ends up balancing as the season goes on, then that can be waved off.
Thinking along the same lines - more explosive players in the East coming thru and scoring to finish games thru weaker D teams?
Again may prove false if numbers end up balancing out... but presently 3 vs 15 is huge difference.
FANonymous
Registered User
- Nov 7, 2010
- 4,911
- 0
It means teams in the West are tired from all the extra traveling they have to do and therefore are unable to put forth as much effort offensively as the East squads do, thus resulting in more OT and SO games.
Feed Me A Stray Cat
Registered User
CanadienShark
Registered User
- Dec 18, 2012
- 37,554
- 10,826
You can apply an "if" statement to anything though. If the East becomes lower scoring and the West becomes higher scoring, then perhaps it would change.
Feed Me A Stray Cat
Registered User
DL44
Status quo
??
You must be one of those guys in a power rankings thread that posts a link to the standings... (second time i've used that line in this forum.. gotta get new material)
Just a speculative discussion.... those usually include 'if's' and 'what if's'. You new at this?
Last edited:
LTIR Trickery
Plz stop pucks
Its basically this at this point. Its an anomaly within a small sample size.
ChadSC
SportingCharts
It's very likely just randomness.
Looking at past shootout data (2005-2011 seasons) there have been a total of 1,138 games that went to the shootout.
Western Conference teams have been in 1,163 shootouts while the Eastern Conference teams have been in 1,113 games. As a percentage: West = 51.1% and East = 48.9%.
Looking at past shootout data (2005-2011 seasons) there have been a total of 1,138 games that went to the shootout.
Western Conference teams have been in 1,163 shootouts while the Eastern Conference teams have been in 1,113 games. As a percentage: West = 51.1% and East = 48.9%.
Yossarian54
Registered User
As of now:
East; 140G, 400GF, 2.86GF/G
West; 140GP, 376G, 2.69GF/G
So while there is a small difference in total goals scored, I don't think it's large enough to explain a 5:1 ratio of shootouts.
However, if you look total the absolute values of the GF/GA differentials you get:
East - 108
West - 86
Which is a much larger relative difference than total goals scored between the conferences, perhaps suggestive of 'tighter games'.
However, it may actually be a result of Tampa's dominance of the SE, as they are the worst (best?) offender in the GF-GA totals. Likewise WPG, WSH and FLA are the next worst (worst.) offenders. So 22 out of the 108 (absolute value) total goal diff is due to Tampa and the SE, which has only been 4 games out of 140 - not really enough to influence the shootout stats.
TL;DR: Not scoring.
Unless someone wants to dig through every game for each conference and plot the goal diff, as there could be counteracting effects.
I think to solve this you need to simplify the problem.
Understand the difference between the team with the most shootouts to the team with the fewest. The conference is likely a sum of this understanding.
Understand the difference between the team with the most shootouts to the team with the fewest. The conference is likely a sum of this understanding.
DL44
Status quo
Presently...
East: 8 Shootouts
West: 21 Shootouts.
That's a few days past the quarter season mark.
Sample size is more significant and the conferences are still trending very differently with the west having almost 3 times the amount of shootouts.
Will continue to monitor.
East: 8 Shootouts
West: 21 Shootouts.
That's a few days past the quarter season mark.
Sample size is more significant and the conferences are still trending very differently with the west having almost 3 times the amount of shootouts.
Will continue to monitor.
Master_Of_Districts
Registered User
It's still random variation.
There's no reason why either conference would systematically have more shootouts than the other.
me2
Go ahead foot
It's likely randomness but I would like to see the average goals scored by the winner and average scored by the loser. I imagine the west has a lower winners goals scored avg and a higher average for the losers. (or just a margin)
A anomaly is normal! 
It really is. If there are like what, 500 statictical posts -- it would be extremely unnormal if one of them didn't stick out.
I read a very fascinating article on this in the NY Times a couple of years ago. There was a professor in the US who had done a case study I think on the death of a two twins in Finland. If anyone remember, there were some talk about a group of people someway connected to the defense industry in the US that died from accidents right after 9/11, and there were some speculations as if they had been assassined. Anyway, in connection with this, this professor commented on his research.
Two males in a small Finnish town, around say 85 y/o, died in separate car/pedestrian accident on the same hour the same day (like withint 30 minutes of each other or something like that). Subsequently, the police went out and said something like "this must be a murder, its just too unlikely that two guys who lived for 85 years then is killed within the same hour in separate accidents. One pedestrian in this city is killed in a car accident every third year, we can almost rule out accident". But it subsequently bascailly proved to be accidents.
This professor from the US started like with a number of all identical twins in Finland. He got a number of like a couple of millions.
-Then he looked at the number of car accidents in Finland and got a extremely small percentage for two identical twins to be killed on the same day.
-Still it seemed extremely unlikely that this would happen. But he found out that this winter day in Finland was extremely icy. So accidents was alot more likely on a day like that. The percentage went up a bit.
-He also found out that both twins had the same medical conditions. They were a tad demented, had very poor hearing and didn't see that well. The percentage went up even more.
-He found out that both twins as a routine went out around the same time the same day, which cut the number back more (if they were to die in a accident like this, it were on that day of the week at that time).
-Then he looked into other forms of accidents in light of the above. And the percentage for two identical twins, with the same medical condition, to be killed on the same day went up alot.
-Then applied that percentage to all identical twins in Europe. And the percentage for it to happend went up alot (he went from say 60k identical twins in Finland to 2m in Europe).
In the end he could basically prove that how if you looked at like a 20 year period, an accident like this were to be expected in Europe. Beyond that, he also noted that this probably was one of tens of thousands of accidents that would have been as newsworthy. Ie, if you just don't look at identical twins, but like accidents could happend were grandmother, mother and daughter to be killed in three seperate accidents on the same day -- that would also be a world wide story. Two sibblings being hit by the lightning. Falling airplane or whatever. His point was basically, if you live a full life you will experience a whole bunch of extremely unlikley things.
It really is. If there are like what, 500 statictical posts -- it would be extremely unnormal if one of them didn't stick out.
I read a very fascinating article on this in the NY Times a couple of years ago. There was a professor in the US who had done a case study I think on the death of a two twins in Finland. If anyone remember, there were some talk about a group of people someway connected to the defense industry in the US that died from accidents right after 9/11, and there were some speculations as if they had been assassined. Anyway, in connection with this, this professor commented on his research.
Two males in a small Finnish town, around say 85 y/o, died in separate car/pedestrian accident on the same hour the same day (like withint 30 minutes of each other or something like that). Subsequently, the police went out and said something like "this must be a murder, its just too unlikely that two guys who lived for 85 years then is killed within the same hour in separate accidents. One pedestrian in this city is killed in a car accident every third year, we can almost rule out accident". But it subsequently bascailly proved to be accidents.
This professor from the US started like with a number of all identical twins in Finland. He got a number of like a couple of millions.
-Then he looked at the number of car accidents in Finland and got a extremely small percentage for two identical twins to be killed on the same day.
-Still it seemed extremely unlikely that this would happen. But he found out that this winter day in Finland was extremely icy. So accidents was alot more likely on a day like that. The percentage went up a bit.
-He also found out that both twins had the same medical conditions. They were a tad demented, had very poor hearing and didn't see that well. The percentage went up even more.
-He found out that both twins as a routine went out around the same time the same day, which cut the number back more (if they were to die in a accident like this, it were on that day of the week at that time).
-Then he looked into other forms of accidents in light of the above. And the percentage for two identical twins, with the same medical condition, to be killed on the same day went up alot.
-Then applied that percentage to all identical twins in Europe. And the percentage for it to happend went up alot (he went from say 60k identical twins in Finland to 2m in Europe).
In the end he could basically prove that how if you looked at like a 20 year period, an accident like this were to be expected in Europe. Beyond that, he also noted that this probably was one of tens of thousands of accidents that would have been as newsworthy. Ie, if you just don't look at identical twins, but like accidents could happend were grandmother, mother and daughter to be killed in three seperate accidents on the same day -- that would also be a world wide story. Two sibblings being hit by the lightning. Falling airplane or whatever. His point was basically, if you live a full life you will experience a whole bunch of extremely unlikley things.
If it was purely random we could test the null hypothesis. I mean, we could find out if the variance was due to chance alone. But I don't want to do some data entry and then run a t-test, and I don't think I have to. Here's why.
On the surface, it looks big. I see about a 3:1 ratio from West to East (21:8) and the West plays a different hockey game than the East. But no teams face each other from different conferences, hence an OT game in a conference perpetuates this ratio.
I then see a team like Nashville having 8 OT games out 16 and then see a team like Toronto have 1 OT game out of 16. Is there something about these two teams that could explain the difference in games that went into OT? Possibly, they play two different games really.
So I dug a little deeper....
In between Toronto & Nashville are 28 other teams that vary from 3 OT games to 6 games then went into OT. No Western Conference team has less than 3 OT games, and no eastern conference team had more than 5.
That sounds legit, right?
Well, hang on.
6 teams went into OT 5 times (only 2 were Eastern)
4 teams went into OT 4 times (2 were East/West),
7 teams went into OT 3 times (only 4 were Western)
6 teams went into OT 2 times (only 1 were Western)
There's a lot of teams hanging around the median (3) and the mean (3.6). Hence the distribution should look pretty close to a bell curve. So the spread is normal and there's some overlap between the conferences.
I think we should look at this again at the end of the season to see if that ratio has reduced, but so far I would agree that this is likely a coincidence given the sample size.
On the surface, it looks big. I see about a 3:1 ratio from West to East (21:8) and the West plays a different hockey game than the East. But no teams face each other from different conferences, hence an OT game in a conference perpetuates this ratio.
I then see a team like Nashville having 8 OT games out 16 and then see a team like Toronto have 1 OT game out of 16. Is there something about these two teams that could explain the difference in games that went into OT? Possibly, they play two different games really.
So I dug a little deeper....
In between Toronto & Nashville are 28 other teams that vary from 3 OT games to 6 games then went into OT. No Western Conference team has less than 3 OT games, and no eastern conference team had more than 5.
That sounds legit, right?
Well, hang on.
6 teams went into OT 5 times (only 2 were Eastern)
4 teams went into OT 4 times (2 were East/West),
7 teams went into OT 3 times (only 4 were Western)
6 teams went into OT 2 times (only 1 were Western)
There's a lot of teams hanging around the median (3) and the mean (3.6). Hence the distribution should look pretty close to a bell curve. So the spread is normal and there's some overlap between the conferences.
I think we should look at this again at the end of the season to see if that ratio has reduced, but so far I would agree that this is likely a coincidence given the sample size.
Last edited:
DL44
Status quo
Also nullifying the difference between conferences is you should look at the total number of shootouts as a percentage of games played instead of straight up...
i.e.
if the west has 21 SO's in say 160 games, and the east with 8 in 160...
it's 13.1% vs 5%... a variance 8% between the leagues... as opposed to looking at it like almost 3 times the amount.
So what amount of variance can be considered significant at the end of the yr? Is it 8%? 5%? 10%?
At the end of the yr, a 5% difference is an additional 16 SO gms for one conference over the other... 10% is 36.. 8% is a 29 gm difference.
Maybe if we lay out what would constitute a significant difference, we can then gauge how the trends are unfolding as the sample size continues to increase.
Until suggested otherwise by stats peeps.. i think using a 10% difference at yr's end as marking a significant difference would be a conservative start.
.........
Presently...
East 117 gms: 9 SO's, 7.7%
West 115 gms: 23 SO's, 20%
(12.3%)
To me.. right now.. for that particular sample size (232 gms), the difference appears to be a significant one.
I also understand that there 488 more games to be played, so we will see.
Update Feb 25th morning:
East: 136 gms. 10 SO, 7.4%
West: 132 gms: 24 SO, 18.2%
(10.8%)
So another week.. and another move towards each other.
i.e.
if the west has 21 SO's in say 160 games, and the east with 8 in 160...
it's 13.1% vs 5%... a variance 8% between the leagues... as opposed to looking at it like almost 3 times the amount.
So what amount of variance can be considered significant at the end of the yr? Is it 8%? 5%? 10%?
At the end of the yr, a 5% difference is an additional 16 SO gms for one conference over the other... 10% is 36.. 8% is a 29 gm difference.
Maybe if we lay out what would constitute a significant difference, we can then gauge how the trends are unfolding as the sample size continues to increase.
Until suggested otherwise by stats peeps.. i think using a 10% difference at yr's end as marking a significant difference would be a conservative start.
.........
Presently...
East 117 gms: 9 SO's, 7.7%
West 115 gms: 23 SO's, 20%
(12.3%)
To me.. right now.. for that particular sample size (232 gms), the difference appears to be a significant one.
I also understand that there 488 more games to be played, so we will see.
Update Feb 25th morning:
East: 136 gms. 10 SO, 7.4%
West: 132 gms: 24 SO, 18.2%
(10.8%)
So another week.. and another move towards each other.
Last edited:
DL44
Status quo
East: 161 gms. 15 SO, 9.3%
West: 157 gms: 28 SO, 17.8%
(8.5%)
So a continued regression towards each other...
West: 157 gms: 28 SO, 17.8%
(8.5%)
So a continued regression towards each other...
DL44
Status quo
East
277 gms, 34 SO.. 12.3% of gms
West
272 gms, 43 SO.. 15.8% of gms
Down to a mere 3.5% difference between conferences.
So a definite regression towards the mean to point i think we can state the differences between the 2 conferences is no longer significant.
If it remained at around 10%, we would of had something more substantial to discuss regarding real differences in each conference this yr. oh well. will do a final tally at season's end for those interested..
277 gms, 34 SO.. 12.3% of gms
West
272 gms, 43 SO.. 15.8% of gms
Down to a mere 3.5% difference between conferences.
So a definite regression towards the mean to point i think we can state the differences between the 2 conferences is no longer significant.
If it remained at around 10%, we would of had something more substantial to discuss regarding real differences in each conference this yr. oh well. will do a final tally at season's end for those interested..
Ad
Latest posts
-
Series Talk: WCSF - Vancouver Canucks vs Edmonton Oilers- Latest: LickTheEnvelope
-
-
-
Speculation: 2023-2024 General Lightning Discussion - Part 3
- Latest: FinnLightning26
-
