You do not need that many people to have VERY accurate statistics.
A year or so ago, a local community in my area was told about a new developement in the area, they were told it was going ahead because their "polls and statistics" said that a large percentage of people wanted said developement.

The resulting outrage, fury and MASSIVE objections from people WHEN THEY HAD ACTUALLY BEEN ASKED if they wanted it or not was enough to cancel said project. Their "statistical polls" said it would be a welcome addition to the area.

They got the shock of their lives when said community, once they knew of this little concept and its intended implementation, stood up and said HELL NO get this out of here, we refuse and we will protest and fight you on it.

Statistics ( and yes I do know the concepts and the ideas of margins, polling levels, sample sizes etc ) can be useful. They can also be an avenue for a huge mistake when you take a miniscule number of peoples opinions as gospel.

The matter was then quietly and VERY quickly dropped. They polled maybe 100 people out of a community of several thousand, and took said poll results as proof that the majority approved of the idea.

THEY. WERE. WRONG.

Statistics, when done properly, are a useful and at times essential tool for gauging the probability of success or in determining trends in a specific area. Their major weakness is that said figures can then be manipulated or misused or misread....in the end what you get is, guess what. GIGO.

If we wanted to determine the health of the game I feel an easier way to do this is just to wait for their financial reports, rather than guess at something that could be business as usual or misleading us given some interesting relationship we're not seeing because we don't have a good full picture.
On this we agree, with a caveat. I refer back to WOW, and their trumpeting of their "record sales" and their "successful expac"...leaving out the fact that a few months in , all those new customers walked out the door in disgust.

Their "statistics" said the game was a success. Their horrendous attrition rate, was, however, glossed over or conveniently omitted.

An MMO lives on money, agreed, at the same time is it just as valid an evalulation of said MMO when they consistently and pathologically fail to retain said customers. Acquisition of customers is a "loss leader" as it costs far more to gain new customers than it is to retain them.

I stand by my previous comments on sample sizes, and from my POV, polling 0.2% of a large population is idiotic. 0.2% does not in any way accurately reflect reality. A local council and a specific construction company lost 2.1 million dollars as a result.....which then bled over into the next local council elections, which resulted in those responsible being turfed out on their ear.

Yeah. Ouch.