Its important when you look at the game from the perspective of the current meta-game. I know some people prefer to take a more objective approach and context is still important, but both are relevant. Thatd actually be an interesting discussion too.
Not to brag, but I predicted a lot of things back in 09 based on my intuition and (limited) understanding, and I'm not even a good player.
Back then I thought MK Olimar and ICs were the best chars in the game. I thought Wario, who was ranked like 3rd at the time, wasn' that good, and that Sonic was capable of a lot more in reality than what people were theorycrafting. I thought Pikachu was good character who could be a threat on a national level and to high-top tiers like MK. There were other little things too but I won't go into them.
The tier list is starting to reflect this now. The point isn't that I'm some amazing authority on Brawl, in fact it's the opposite. The fact that a low-level player like could predict developments like that years in advance shows how limited the results-based system is, even if it's not entirely results- centred.
It's much more accurate to look at things like frame data, human reaction time, learning curves etc. than to base a character's potential on whether they have a good main at the time or not. Because that's the thing, a character's potential isn't based on how good its players are, but it's actual programming.
Also, of people want to do MU ratios, it'd be much more pragmatic to do ratios that apply to a level of play that more than like 4 people are currently at.
People say things like "DK experience doesn't mean crap unless it's against Will" well then what's the point of an MU ratio or MU advice if it only applies to one person in the world? How does that help the 99% of the Smash community who play at a level that these ratios and tier lists apply to?