My problem with that is it becomes very difficult to isolate all the various permutations of what they claim to have found. If I am reading it correctly, CD's rank about "100", SACD's and 24/96 PCM about 180, "upconverted" CD's about 170, 320k MP3's about 50, and it appears that each cycle of WAV > FLAC > WAV costs you about 10 points (or maybe 15) on this scale. So it only takes about 5 of those FLAC conversion cycles to reduce a CD to 320k MP3 quality, and about 8 for a 24/96 download to be equivalent to 16/44.1 CD. Oh, but wait, if you are using Jplay for regular 16/44.1 CD-quality files, all of a sudden you're up at 150 again, so it only takes about 4 FLAC conversion cycles for a 24/96 file to be worse than the CD.
Are you beginning to see some of the problems here? Not to mention the fact that if 1000 FLAC conversion cycles produce no change at all, what does that mean about the rest of their claims?