I know what sample size is.
Its nice to see stats guys admit that hockey stats are no better than flipping a coin.
Then you have trouble reading. Nowhere in my message did I imply that hockey stats were as useful as flipping a coin.
In case you didn't get it the first time, it's very simple: one game of hockey is flukey (I said 'one', not 'the').
Some anti-stats guys like to say it's because hockey is chaotic, cannot be predicted, look there: momentum and other stuff like that. While there is some truth to this, it's not the complete story.
The major problem* with ONE game of hockey is sample size. 30 shots and 2.5 goals per team is not enough for stats. Just like flipping a coin 4 times is bound to be flukey and not representative of the real, underlying physical phenomenon you're trying to assess.
You get it now? A perfectly predictable and deterministic process (coin toss) becomes totally unpredictable and random when you don't sample it enough. So yes, ONE game of hockey is inherently unpredictable, but that's because of sample size first, not because of the nature of the game.
Anyways, I should know better than to argue with the anti-stats crowd. After all, hockey is a very conservative sport...
Just kidding, guys. Go Habs and yes, I'm hoping for a flukey result tonight!
*Note here that I say 'problem' in the context of statistical analysis. A 100 hours hockey game, if it was possible, would be great for a stat geek, but extremely boring to watch. First, the result would become much more predictable, and we'd all be dead drunk by the 14th period.