I don't see why, exactly.
Again, "check the tourney results" is a kind of bogus answer. Which tourney results? The ones reported in the Sonic boards thread that are only reported when Sonic cracks top 10, or the ones reported in Ankoku's thread? Either way, neither of them can really be called statistically relevant.
I've done a lot of thinking on this, and to me, the only tournament results that should even matter are large regional, national, or international ones if results are even to be counted at all. For instance, Pit barely places at all anywhere that matters, but it isn't because Pit is a bad character, it is because there are no really amazing Pit players outside of Japan. The SBR even recognizes this to an extent, because Pit is thought of as being a pretty decent character; 16th or 17th best, in fact.
The best way to create the tier list is to have these guys come out with an official, thought-out match-up chart. Then, add the first number of each character match-up ratio together and add weight based on top and high tier match-ups. For instance, huge **** vs MK matters more than huge **** vs. Peach. If they want to factor tourney results, for instance, with Dedede actually working out despite him not having the same amazing match-up spreads as the other top tiers, that's fine, but they should not be factoring Joe's Smashfest Tourney that included 13 people no one has ever heard of. Large regional tourneys are much easier to track. It is easier to make sure the rules are "SBR approved," it is easier to make sure that a certain calibur of player is attending, and most importantly, there are more people attending, meaning the sample size per tournament increases dramatically. It's pretty easy for an amazing CF player to beat 13 nobodies, for instance.
You could argue whether or not match-up ratios are canon. That is a valid response to my idea, although I think the way we talk about match-ups is mostly fine, assuming both parties are intelligent contributors.