Match-up charts are simply impossible to make work right. There are 2601 match-ups in this game, a good number of which have probably literally never been played out in tournament (find me a Mii Swordfighter vs Mr. Game & Watch tournament match) and most of which have been played sufficiently rarely that we really don't know much.
Even if we somehow were able to play each a lot, we'd run into ruleset asynchronization. Stage rules vary region to region (or even locality to locality), and this can make a big difference since I'll tell you from my experience of a MU I know well that Sonic vs Rosalina is completely different if we're playing on Final Destination, Smashville, or Battlefield just to stick to stages everyone is familiar with and the nuances continue to mulitply as we see how it plays out on a stage like Wuhu Island or Kongo Jungle 64 which are legal stages in my region but not others. Custom rules also vary; the game is just too different if customs are not legal in an area to use the data they may gather (you can't just "add in" customs to your thinking at the end; they're too fundamental). Speaking of customs, remember that players are going to need substantial time to figure out which particular custom load-out is optimal in each match-up which may not always be obvious (like I still debate with myself over whether 2311 or 2211 is best for Rosalina against Sonic, and I've played that MU over 100 times against Sonics good enough to make me care).
If we somehow solve that, we also have the fact that there is no real scale. In theory a match-up ratio is a probabilistic statement. If two high level equally skilled players play 100 games, the ratio shows us how many each will win. In reality this is a myth. Two equally skilled players don't exist. High level players for every character don't exist. 100 games will never be played to measure. Even if they were, player skill is non-static and if your players don't suck they'll be learning as they go over 100 games and the early games won't be representative of the metagame after 100. The numbers people give are in reality pulled out of their rear ends. We say 60-40 for "kinda bad but winnable" and 70-30 for "really bad". We say 90-10 for "essentially impossible" even though in reality a 90-10 match-up probably cannot exist (any match-up bad enough for a 90% win rate is probably bad enough for a 100% win rate). These numbers are entirely subjective, and what precisely one set of numbers means will vary person to person. If I talk about a 65-35 match-up, just how bad is it? I bet I can get at least five meaningfully different answers. Of course, within anyone's region, you'll see other biases; if your local Diddy player is your most skilled player and your second best player is a Sheik main who just can never beat your best player, you might believe Diddy beats Sheik badly when the real reason this loss keeps happening is that the Diddy main is the better player. Even if Diddy does beat Sheik, your view on the margin of victory will inevitably be distorted by the skill gap that has the Diddy main seemingly effortless two stock this Sheik who beats everyone else. Put it all together, and the numbers people give for match-ups are consistently completely meaningless.
Even if somehow you could solve the asynchronous rulesets and the fact that we have no meaningful scale, we'd STILL have to trust the subjective judgment of hundreds of people about game balance. Let's be real here; do you really think we can appropriately vet that many people for being competent at analyzing game design (which is what balance discussion is)? Even being a good player isn't necessarily good enough; often good players understand the game very well intuitively but remarkably poorly when it comes to anything they have to explain (and on the other side, some of the people who can accurately explain how the game works the best are not notably good players). With a small crowd you can carefully vet people. With the hundreds you need for a MU chart? It's impossible; totally incompetent people will be averaged in and it will corrupt your final results to a very real extent. Even beyond the total incompetents, you have to fact that competence is a scale and without a doubt some characters will be considered by smarter people than others, resulting in an uneven distribution of understanding of the game throughout the chart.
Some of these problems still exist when you just make a tier list directly, but the effects can be understood and they're on a small enough scale you can minimize them. A match-up chart is just such a large project that the errors introduced by all of the above factors become unavoidable to such a large extent that the degree to which the errors corrupt your results significantly outpace the degree to which considering match-ups improves your accuracy. A match-up chart based tier list kinda works for games that has been out for 5 or more years with no variables other than character selection and a smallish cast (20 or fewer characters). For modern fighters especially including this one, it's just so thoroughly impossible. It is nothing but wise that we don't try; it would just be us working really hard to produce a non-useful result.
.
)