Tobias - think my last note was submitted before your note was posted, certainly didn't mean not honest as in not truthful.
Anyway, there is an old saying that I like, "In God we trust, everyone else needs data".
An honest blind test would be something like;
1. You have someone connect cable x in your system and write that down.
2. You listen & write down whether you think you are listening to cable x or cable Y.
3. You walk away and have the other person flip a coin, heads they swap cable x for cable y, tails they don't.
4. You repeat steps 2 and 3 enough times to be comfortable with the results (random chance says you will be right 50% of the time so you want a few reps. to take away the chance of being fooled by probability). Someone who remembers the statistics formulas could give you an estimate on the probability of accuracy based on sample size, if anyone does the test maybe we can get a statistician type to chime in.
5. If you are right 100% of the time, it would be hard for a reasonable person to argue with you. If we really wanted to be anal about it, the same kind of test should be independently verified once or twice, but since this is probably less important than cold fusion I say that is overkill.
Best Regards.