In terms of putting us through a level matched bias controlled test - I'm all game for that. I would like to know if we're crazy or not. Maybe someone can help me design a test as we intend to keep testing.
You aren't crazy, even if there is no real difference. Deluded, maybe, but that's entirely normal. Humans are like that. Hey, maybe there's a difference. If so, I'd bet one of the cables is damaged or faulty in some way. It's of course possible something is wrong with a cable. If you think something is wrong with the monster cable, get some cheap 14AWG from Radio Shack or something. It will equal the Mega cable, I'll wager.
The following isn't good enough to pass the Randi Challenge, but it's a good start.
Make sure it isn't possible to see which cable is hooked up from the listening position. Draw up a random list (flip a coin a bunch of times, use the random # function in Excel, whatever) of speaker wire A, speaker wire B: ABBBAABABBBAAABA, etc. Make sure there's an equal number of A's and B's.
Choose a test sample where you think you can hear a difference. Go ahead and listen in "un-blind" conditions at first.
Then, One person is in charge of switching the cables (or pretending to switch the cables). Take whatever steps you think are necessary for the other person not to know what switch (if any) is taking place. One person listens, and based on what you heard at first, take a guess which wire you are listening to. Keep track, but don't communicate about this AT ALL until the end. Don't let the person know if they're right or wrong. You shouldn't even know what they're writing.
You may then want to switch roles (using a different random list, unknown to the new listener).
The more trials you do, the better. With 20 trials, 15 or more correct is likely "real", anything less is usually considered "chance". With only 10 trials, you'll shouldn't accept anything less than 9/10 as "real".
see:
http://www.provide.net/~djcarlst/abx_bino.htm
(The usual acceptable limit is p=.05 or less is considered "real")