Six degrees of separation debunked?

The latest reports flying round seem to be saying that "six degrees of separation" theory has been "debunked". When I read the original BBC story I have to say I was less than convinced. Basically the new chatter is based on the fact that theory was tested by getting people to try and communicate to a distance person via a chain of mailed letters, and that very few of those letters (less than 10%) ever reached their target.

Big deal.

I'm going to call Foul! on this supposed debunking because the test of the theory was a pretty lame one and prone to errors. For instance what if the person just doesn't bother to send the letter? My high-school statistics seems to indicate that if just 1/3 of people receiving a letter didn't bother to mail it on then with a chain of six that would yield only a probability that only 8% of letters actually reached their target.

More important, think about if someone was connected to another by one and only one chain of six people - the test presupposes that each person in the chain knows the ideal person at each step to contact. The likelihood of that being true is unimaginable. What actually happened, I expect was people just forward to someone in another city closer to the target person and by chance the connection was found - there was certainly no rigorous exploration of connections. All things considered a success rate of 10% or even less seems very high to me and I would actually guess that the results point to people being even more connected, on average, than one might imagine - maybe only three or four degrees in many cases.

I hope someone else is able to point this out - I doubt if anyone will notice my protests. But perhaps by the theory of the fundamental interconnectedness of all things mine will reach someone with enough clout to quash the debunking theory. As far as I'm concerned there are people out there like Google, Friendster, Yahoo and others (even phone companies) who, by using their databases to analyze social networks and patterns of communication could easily measure with much greater accuracy the n-degrees of separation. In fact I would be willing to guess that the CIA and others have already done this - they just aren't telling us about it yet.