I propose that the same can be said for any character (Including Pichu.)
Of course, that's actually the reason why match-up discussions are generally meant to be deductive, to avoid that issue. Of course, this runs into the problem of the lack of a good theoretical framework, but we're working on that.
The the thing is, tiers exist, because it's impossible for any two things to be exactly the same. The difference may be minor, or the difference may be major, we're just trodding along to find it as best we can.
I guess my biggest question to the tier-ists out there is thus: Has there ever been any research conducted under reasonable experimental conditions to suggest that any one character might prove to be consistently better than all of the others.
Tournament evidence is the best we can do.
1. All characters are controlled by the same level CPU
...
No, CPUs don't abuse the tactics which make the characters good in tournament play. In melee, they almost never wavedash, I've only seen them dash-dance like once, I've never seen them pull off a ton of the combos and they never camp, ever.
Forget SHL.
The same applies for Brawl, when's the last time a cpu chaingrabed you to a dair spike with Falco?
3. There only stage used will be Final Destination, as it is regarded as the most neutral stage.
...
No, FD is actually one of the less neutral stages due to the screwy edges which hurt some recoveries a lot, the ungodly blast zones, and the power it gives to camping.
Using this formula you could accurately discern after repeated trials (I'm talking hundreds of tournaments here) whether or not any one character was better than the others by simply tallying up the results.
Not at all. It would tell us how the game would be if we all reacted like computers and didn't do any ATs.
I would like to use regular tournament play in such a study, but because of the high degree of variables and the large human element to said tournaments, they simply cannot be included in any empirical research.
It's better then simply cutting out most of what makes the tier list.
After its all said and done, I'm on the side of the data. If anybody can show me a study bearing any sort of passing resemblance to the one I have outlined above, I will spin on a dime and proclaim tiers to be true. Until then, I remain unconvinced.
Then I welcome you to the "problem of induction", look it up.
Deduction is the only truly reliable way to obtain a conclusion.