View Single Post
02-19-2013, 09:52 PM
Registered User
Join Date: Nov 2010
Country: Canada
Posts: 12,514
vCash: 500
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.

Last edited by MarkGio: 02-19-2013 at 10:20 PM.
MarkGio is offline   Reply With Quote