This is why I think every single fan poll is inaccurate. Why remove all the Krystal votes? What makes her different from Isaac, other than Isaac is popular here? Same thing happened with one of those reddit polls where people set up bots to vote for Daisy and King K.Rool. Everyone on Reddit was quick to toss out the Daisy support, but insisted K.Rool's vote count was fine. Because K.Rool was popular there, and Daisy wasn't.
Even if K.Rool is the most popular character everywhere, which I happen to think he is, the polls are still right for the wrong reasons by continuously polling the same insulated community that has been shown to spam votes for him over and over again. Even the most visible fan polls hardly return a fraction of the votes the official ballot has received, and while a good sample size for this kind of poll only needs to be a small fraction, usually between 3 and 5% to be accurate within a 5% margin of error, that only applies if the poll covers the entire voting base.
Poll 1000 voters from around the US about who they'll vote for President, you'll get not too terribly far off from a 50/50 Democrat/Republic split. Poll 1000 voters from around the US, but 700 of them are in California, and you'll wind up with insanely skewed Democratic results. And maybe this was 2012, where the Democratic candidate won. Does that mean the results were accurate? No, because they'd have been exactly the same in 2004, when the Republican won, because California always votes democrat, just like Reddit always votes K.Rool. I'd be quite interested to know whose popular in a poll that didn't include Reddit or Smashboards.
The polls never show us anything we don't already "know" because the results are manipulated to show what we think is true.