Hey guys I just wanted to share this little “case study” I did with you because I think a lot of people may find it interesting.
It started from a discussion me and a friend had. I was bored one day and thought I would work this out just for the hell of it. I mean, I spent 5 years getting a college degree may as well use it right? 
there’s been a thread posted about this (at least the sets part) before: The real reason why top-tier always wins in america
as you can see, it didn’t go well since srk is full of idiots
nice to see it presented in a more formal manner though and the tournament stuff is enlightening. good work
Nice work, but I think you have to be a little bit careful about taking match-ups ratios as forward looking probabilities. Match ups lists basically tells you what the AVERAGE win ratio is at a given time. The actual win ratio varies from player to player, also there is no guarantee that the win ratio would stay the same forever.
Of course, you could put in the assumption in your analysis that all players are perfectly rational, who know everything there is to know in the game, and only make the best judgement at any given time, to make it valid. However, we all know we are not perfect, unless you work in some Ivy league econ department.
p.s. I’m a grad student too, can’t wait to join the phd fighters club!
Funkdoc posted this a few years ago. edit: beaten I cannot into statistics so I just hit buttons and hope for the best.
That is why I had the segment at the end saying that you can also look at this not just as character match ups, but your own match up against either players or specific characters. I mean even if a match up is 5-5 if you have trouble with it then your odds of winning can drop or if you are playing someone who has trouble against you then your odds of winning go up.
If we wanted good hard probabilities we would have to actually do the stats on the match up of interest.