I guess the question is, What is it in NPD's formula that conclusively biases the results, and makes one's numbers more wrong than another?
Sort of like Nielsen TV ratings. People complain because the sampling is so small, but results are generally "equally wrong" and therefore are at least relatively accurate ratio-wise among each other, even though exact figures are elusive.
So what is it that NPD does wrong that slants the results? Is it Sony's larger retail presence in general (more stores = more track samples)? The world may never know...?