October 8, 2007

Average Predicted Standings



NamePL
U
DET231
ANA361
S.J461
VAN591
CGY591
MIN693
COL6101
DAL7113
NSH9135
STL10128
L.A11138
EDM121510
CHI131511
CBJ131510
PHX151514


NamePL
U
OTT251
PIT241
NYR351
BUF471
CAR7141
N.J7113
TOR8124
T.B9162
FLA9153
PHI10145
ATL10154
MTL10156
NYI12156
WSH13159
BOS141511

I like NHL standings predictions, because I find it interesting how wrong we can be. The above standings represent the average (P) of several different standings I found on the web (if you know of more I am more than happy to add them).

I have also included the minimum standing spot (L) and maximum standing spot (U) based on the variance of the predictions. For example the New York Islanders have had a lot of different predictions (from average to great to terrible) as such they have a very large expected range (anywhere from 4th to 15th), where as Phoenix has a very small range (15th). No one has predicted Phoenix to do better than 15th.

The above standings is the average of several different sites:
Bookies
Mirtle
Mirtle's Playoff poll
My opinion
McKeen's Hockey
The Hockey News
Added:
Alain Chantelois
Gaston Therrien
Howard Berger
David Johnson

10 comments:

David Johnson said...

If no one predicted Phoenix to finish anywhere other than 15th, how can Columbus, Chicago and Edmonton also have an L of 15? That doesn't make sense.

JavaGeek said...

The L and U are based on standard deviations and with only 6 predictions they become large very quickly.

If a team was only predicted to finish 7th and 8th (3 for each with an average of 7.5) would have a variance of 0.3 a standard deviation of 0.55 so the L = 7.5 + 0.55*2 = 8.6 and P = 7.5 - 0.55*2 = 6.4. No one picked 6 or 8, but the range still includes it.

With Phoenix, since the variance is zero it has no range.

So a team who was predicted to be in 14th and 13th would have an L of 15.

Mogen_david said...

Given the nature of the data wouldn't Tukey's 5 number summary or median, 2nd and 3rd quartile range have been a more reasonable summarization than StDev? And if your going to stick with normality a confidence interval. It depends on if your trying to show the variation or the confidence in the estimate (mean). In the end I think I would have gone for min, max and mean over mean and +/- one StdDev. But that's just me.

David Johnson said...

I figured that but with the small sample size it becomes next to pointless to consider. Even just using high prediction, low prediction would get much of the point across. Or else use more predicted standings. I know there are more out there including my own.

JavaGeek said...

"the variance is a way to capture its scale or degree of being spread out"

This is exactly what I wanted. Also MySQL has a STDDEV function built in and basically I'm too lazy to do it any other way.

It's not meant to be precise or "correct" simply there to give an idea how much spread there is for each team (in comparison to other teams).

MIN/MAX: 1,8,8,8,8,15 is very different from 1,3,5,8,11,13,15

Finally,
As I stated: "if you know of more I am more than happy to add them"

Mogen_david said...

StdDev works as a measure of spread for a distribution with normal error. The "odd" values you are seeing is because the values aren't normal. Your truncated at 1 and 15 to start with. The asymmetry of the distribution is causing the answer to unrepresentative of the values and their distribution.

I like your summary by the way; it's just not the way I would have chosen to present the data. No I take that back it is the way I would have done my first take and wouldn't have thought differetnly but I'm refreshing my R skills and just went though some of the univariate summary functions. And, I wish had the time to get my own scrapper and database up and running. Instead I just harass you and the other stat geeks out there.
I did find a bit of SQl to calculate percentiles by the by.

Mogen_david said...

Oh and forechecker put up his picks.

The Puck Stops Here said...

My picks have been up for a while
West Conference and East Conference.

James Mirtle said...

I've actually got Anaheim in seventh, so I guess I was way out to lunch.

Or not.

S. Pace said...

Where's the rest of the goaltending reviews (Rangers, Buffalo, Anahiem, Montreal, etc)?