| T O P I C R E V I E W |
| Beans15 |
Posted - 03/29/2010 : 11:06:02
Mr. Hugh G. Reaction posted a link to a TSN report about a candid NHL poll conducted by ESPN. Here is the story and a few highlights.
Interesting to note that this poll was 50 players deep, which is about 5% of the entire NHL. I don't believe this makes the poll very scientific, however it does spark debate. The most interesting point on this poll is that nearly 40% of the NHLer's polled believe the NHL should contract.
http://tsn.ca/nhl/story/?id=315609
http://sports.espn.go.com/nhl/news/story?id=5028185
(The actual poll, 20 questions deep)
A list of some selected results from the poll:
BEST PLAYER
Alex Ovechkin 64%
Sidney Crosby 24%
Evgeni Malkin, Henrik Zetterberg 4%
Pavel Datsyuk, Joe Thornton 2%
MOST OVERRATED
Roberto Luongo 6%
Vincent Lecavalier, Dion Phaneuf, Jay Bouwmeester, Sean Avery 4%
BEST FRANCHISE
Detroit 60%
Pittsburgh 14%
Montreal 12%
New Jersey 8%
NY Rangers, Chicago 4%
WORST FRANCHISE
Phoenix 28%
Atlanta 18%
Islanders 14%
Edmonton 12%
Florida 10%
WHO WILL WIN THE STANLEY CUP?
Chicago 32%
Pittsburgh, San Jose 16%
Washington 14%
Detroit, New Jersey 8%
And some other highlights:
Smartest Coach: Mike Babcock, Red Wings 22%
Last coach they would want to play for: John Tortorella, Rangers 18%
Grade Gary Bettman should get as commish: 30% said "C"
38% said the NHL should contract
|
| 12 L A T E S T R E P L I E S (Newest First) |
| n/a |
Posted - 03/29/2010 : 19:28:58
Yeah, Hugh knows his stats - I remember hating that class in university, lol. Brutal stuff, and not my bag at all. In the end, you take it as a general opinion that is safe to say is somewhat trustworthy in bent . . . may not be exactly accurate, but it's the general trend.
Like I said - the more I think about it, the more amazed I am by it . . . because contraction, if it's even supported by, say, a third of the players - that's enough of a faction to make the players (and their union and reps throughout the game) not a negative force against it.
I am excited about the possibility of contraction . . . weird, but true. I think as well, that the players have a much closer and more realistic view of the financial situation in the states especially . . . and you have to think they have contraction as an option because of the reality of the situation down there.
"Take off, eh?" - Bob and Doug |
| Hugh G. Rection |
Posted - 03/29/2010 : 16:37:58
I was only assuming 95% because that is the standard for most stats Can studies, reports you hear in the news, etc. If they wanted a 99% interval they would need ridiculously more than 50, but most recognize that 99% threshold isn't desirable for a study such as this. Also, stats is indeed not algebra, and your example of 1/13 doesn't make any sense.
Also, if you wanted '5' from each team, the sample would no longer be 'random'. 50 out of 660 is actually a very significant sample, as some have noted.
In fact, with a 95% interval and confidence interval of 10, all you need is 96 people to represent a population of 20,000. Stats Can releases reports with less than 1% of the population polled all the time, and there results are generalizable. In this case its 13% of the population. In the end, its just a polling of opinions anyways. |
| polishexpress |
Posted - 03/29/2010 : 16:03:52
Hugh, that's a big assumption about the confidence intervals, that could really screw the math up, couldn't, if you were assuming wrong. (I don't feel like looking up the formulas for stats, they are not in reach, but it seems like you are correct in your math.
Impropriety: the math Hugh did is stats, not algebra, you cannot scale down statistics and expect the same answer, it doesn't work like that.
So just because 50/660 is about the same as 1/13 doesn't mean anything.
Good point about the correlation, though. Unless they would release the actual numbers, there is no way to prove a correlation, so all we are doing is making assumptions. |
| impropriety |
Posted - 03/29/2010 : 14:48:47
Mathematically you might find that 50 people could provide an accurate representation.
Scaling that down, though, could you ask 1 person out of a group of 13 and be accurate within the same margin of error?
I guess the only issue I have with it is the same issue HawkInOilCountry brought up. If each of the first 49 people picked different players for the 'most overrated' title, and the 50th person voted for the same player as the 49th did, that player suddenly becomes the most overrated. Seems like more of a coincidence than a statistical correlation. |
| n/a |
Posted - 03/29/2010 : 14:23:47
Yeah, nice stats breakdown, Hugh - I agree, it's a pretty good sample, actually.
Very, very interesting, in that such a huge percentage of players voted . . . for greater job uncertainty. I am guessing that the real fringe players involved in the poll didn't vote for contraction, but at any rate, even a longer career would be threatened by it . . . even for the greats. It really says something to me.
Thanks for the link Beans.
"Take off, eh?" - Bob and Doug |
| HawkinOilCountry |
Posted - 03/29/2010 : 13:38:39
If Stats canada can claim a survey of 1000 people out of 31 million is 95% accurate with a 1% margin of error then I guess 50 players out of 660 isn't bad. Still I don't like it when 3 votes decided the most over rated player.
I would prefer results from 3 random players from each team, that way each franchise is fairly represented.
The arena wall in chicago should be credited with a goal. |
| Beans15 |
Posted - 03/29/2010 : 12:54:13
Thanks for the stat lesson.
However, even with all of your numbers, 50 out of 660 is a pretty low sample for any survey. I would think that 5 or so players from every team (150 players) would be a more reasonable sample to tell the true story of what the pulse of the league is.
However, for the point of putting some players opinions out there and getting some debate going, 50 is a big enough sample.
I was also surprised to see Luongo as the most over rated but the more I think about it, the more I can see the point even if I don't agree.
No surprise to see Phaneuf's name there either. |
| Hugh G. Rection |
Posted - 03/29/2010 : 12:33:42
Actually Beans, if you have taken statistics during your academic years, you would know that you don't need a large sample size to be representative of a population at large.
Since the tsn article didn't list the confidence levels or intervals, I am going to assume its 95% (almost every report you read in a paper is). With NHL rosters set at 22 players each, thats 660 total players, which means 50 comes in at 13.2% of the league.
From my rough calculations, this would mean the margin of error for the respondents was 10. So it's actually an accurate representation of the league 19 times out of 20, within 10 for each answer. Which means for example that between 54% and 74% thought Ovechkin was the best, between 14% and 36% thought Crosby was the best, etc. If we dropped the confidence interval to 90%, then the margin would be very small, around 4% difference. I would probably tend to think they used 95%, but they didn't say.
So if the players polled were truly a random sample, its an accurate represenation. If 30 of the 50 were polled from one team, then the results wouldn't be as indicative of the overall population obviously.
And I'm done ranting about statistics. |
| Alex116 |
Posted - 03/29/2010 : 11:51:11
nuxfan.... don't forget, Luongo can be considered overrated but still be a top 5 goalie, which i think he is. I think part of it is the perception some have over him winning the gold medal. Back in 02, Brodeur came in and rescued a floundering team to win gold. Luongo's team was already playing well, he just took advantage of an opportunity to take over from Brodeur after Marty's nightmare game. Further, being in Vancouver, the fanfare he rec'd being the local guy prob contributed as well!
A lot of Vancouver fans are so happy to finally have a top goalie there that they're often guilty of over hyping Lu! Mix in a year like he's having (which, for the record, isn't as bad as many make it seem) and i can see him being called over rated.
BTW, with only 50 NHL'ers polled, Luongo's 6% amounts to 3 out of the 50 votes, if i'm in fact doing the math correctly? If so, that's hardly a telling sign.
I was moderately surprised at the love for Chicago to win the cup, but was more surprised at how few picked Washington (compared to what i'd have suspected?)! |
| nuxfan |
Posted - 03/29/2010 : 11:20:38
Interesting poll, thanks for the post Beans. One thing I'd like to know is the conference breakdown for the players - when voting on things like best/worst/dirtiest players, or best/worst teams, biases tend towards players you see most often.
I'm surprised by the number that think Chicago will win the cup. I certainly do not think they will right now, and their chances will rest solely on Huet/Niemi - one of them has to get hot and stay hot thru the playoffs.
Most overrated player Luongo - its interesting to see how Luongo's perceived stock has come down over the last 2 years. When the Canucks acquired him 4 years ago he was projected to be the best goalie in the NHL for the rest of his career - and was the saviour of the Canucks up until they lost to Anaheim in the second round in 2007, where Luongo was coming off his (arguably) best game as a professional and the future seemed bright.
Since then, 3 solid regular seasons (30+ wins, GAA under 2.5, SV% right around .920) but one year of no playoffs and last year's disaster in Chicago. Overrated? I'm not sure, I think he gets the credit he deserves as one of the top 5 goalies in the NHL. Then again, I am biased  |
| Beans15 |
Posted - 03/29/2010 : 11:17:36
Get Smaller. The question:
Should the NHL eliminate teams??
Even with job security at stake, players aren't so sure. Many (48 percent) voted against the notion, but a whopping 38 percent saw its merits. And 40 percent (yeas and nays alike) favored relocating a floundering crew to "Anywhere in Canada."
|
| polishexpress |
Posted - 03/29/2010 : 11:15:14
What does it mean that the NHL should contract? Are they talking about 'contract' in the sense of get smaller, or 'contract' meaning an agreement that you sign? |
|
|
