Oh come on Squilliam, you know what one person's experience is in statistics.  Absolutely meaningless.  Twesterm's personal experience can just be negated by someone else who didn't have troubles. 

I know I am very late quoting this, but I am so glad I don't think this way.  In my life I've owned consoles by Nintendo, Atari, Sega, Sony, Microsoft, and enjoyed them all immensely.

I can't imagine the number of incredible games I would have missed out on had I been "loyal" to a multinational, publicly traded company that has no clue I exist.

Its not one sample. My statistics textbook tells me that I can apply the central limit theorem because I feel that a sample of 80 adequately represents the population of 40GB PS3s at large. The only issue is precisely how accurately he reported the information, so the real problem is not the representativeness of the sample but how accurately twesterm measured it.

And how do you prove this part?  Because obviously if you think it represents the population at large then you can extrapolate the data, but what evidence do you have for that?

Say it!!!

http://en.wikipedia.org/wiki/Central_limit_theorem

"In probability theory, the central limit theorem (CLT) states conditions under which the sum of a sufficiently large number of independent random variables, each with finite mean and variance, will be approximately normally distributed"

A sample size of 80 gives a confidence of roughly 10% for the entire population of 40GB PS3s. So if the sample gives a 10% failure rate we can be reasonably sure it would be between 9 and 11%.

They aren't independent random variables, they're all in the same geographic area, how can you know one bad shipment wasn't brought to where Twesterm lives?

Because the variability of the sample is likely to be small between batches, Sony does their own testing as well to prevent bad batches from leaving the factory. Because the are all produced in the same way using the same methods on the same equipment the expected variability between batches would be small anyway. Im tired, I hardly slept last night and statistics is my worst subject TBH. I can't argue anything stronger than this.

I had a good laugh. Applying a sample of 80 to gain insight into a population of approx. 14-15mill??? May I suggest you re-read that textbook again from cover to cover and then come back and realise what an absolute idiotic assertion that is.

At 99% confidence level the failure rate of all 40Gb PS3s is 10% +/- 8.65 based on twesterns data. ie - could be anything.

Thats even assuming sample data is random and even legit.

See what happens when Kiwis try to read???

Let's not forget the fact that he's probably lying folks.