our research lab did. I don't have the results available any more but there was a fairly "normalised" distribution as I recall. the data was three dimensional in that it was also necessary to take into account line attenuation - noise tends to accumulate in line with line length, simply because the opportunities for noise attraction do so. that's why I burr up when people say xDSL performance *will* deteriorate with distance - it does so only on the basis of an averaged sample. that's great, but Joe Average doesn't think in terms of averages, he thinks in terms of what's applicable to him personally. statistics can only predict the likelihood of that, not dictate it. since 50% of the Joes are going to disappointed by being below the average, and particularly since there's not the ones posting pissed-off "why is my connection slow?" threads, it seems very counterproductive to post a "yes" message when in reality the answer is "statistically yes, but in your case, maybe not". Joe doesn't understand that. if performance "graphs" contained a nice fat band of "normal" ranges of performance (let's say out to about 2SD) from "average" then I would be more comfortable with them. now bear in mind that against a bellcurve, >2SD means < 5% of the total sample. excessive? hell, let's cut it to 1SD - that means over a full third of people are still outside the median... cut that in half (because the ones over are not the ones bitching) and that's still one in every six unhappy. I don't think this makes anyone more serene about their outcome, which is, after all, just statistical. you get what you get. if people stopped ranking themselves in meaningless comparisons the problem would go away. sync speed is not an e-penis competition.