e__
Smash Ace
I am using MK bans as an example continuously in this topic, but this procedure SHOULD be used for any theoretical ban to give concrete evidence as to why something should be banned.
Because there has consistently been talk of having MK banned, if anyone wants him banned they should find numerical and clear data based on facts instead of unbiased claims. In order to best do this, we cannot reject the MK = not broken hypothesis until we have adequate data to prove that he is broken, or that the metagame revolves around him.
First, we need to determine what constitutes the metagame revolving around him, and I think its fair to say that it's generally seen that the characters who do well against metaknight are higher in the tier list and in general get better tournament results rather than most characters who don't do well against him (ie Snake, Diddy, Falco doing better than ROB, Marth, and Olimar). However, this does not disprove the current hypothesis that MK is not enough of a central metagame unit to warrant a ban. I AM NOT SAYING THAT MARTH ROB AND OLIMAR ARE ON THE SAME LEVEL AS SNAKE DIDDY AND FALCO AND I ACKNOWLEDGE THAT THERE ARE OTHER FACTORS THAT GO INTO TIER LISTS.
So, what we want to do is compare the tournament results of a set of MK banned tournaments to the standard tournament results that we have collected recently (as to accommodate for a modern metagame). We should compare the rankings of characters in tournaments without MK to rankings of characters in regular tournaments, and use a Z-Test to see if there is any statistical significance in a character's placing between MK banned and MK legal tournaments (naturally, this will take several months of tournament testing in order to gain adequate data, and hopefully several people with experience in Statistics to collaborate with the Z-tests, determining what is statistically significant/choosing an appropriate alpha-level, sampling, checking work, etc.)
Now once we determine WHO is affected by this test, we need to find out if those characters being affected is enough to warrant the ban. Say only Ganon benefits from MK getting banned, or, in a less radical example, Ganon, CF, Yoshi, and other characters that, even with MK banned, are hardly tournament viable. Is that enough to warrant a ban? Of course not. What about Lucarios, ROBs, Marths, Olimars, and Toon Links? Is 6,7, or 8 characters getting significantly better, or even worse, worth the ban? A limit needs to be determined, say n number of characters, of characters that have statistically significant differences in placings and usage in MK banned tournaments, in order to warrant the Metagame revolving around him.
Ideally, we could determine this n value comparing it to tournaments where a character of similar nature (Snake, Marth, DDD, Diddy, etc.) is banned and testing how many characters are significantly affected by that character's ban, but that occurring is very unlikely, so I hope that should MK banned tests like these actually take place, the back-room determines a fair n value instead of running multiple character banned tournaments (realizing that as less tests take place, much less/more characters will be affected than the true result).
Perhaps even two levels for n characters could be obtained, one for determining whether the metagame revolves around him, and one for determining whether the metagame revolves around him too much.
Naturally, there are multiple ways to do things and I am sure everyone is open for slight modifications of this process, but I suggest we go through procedures like this for determining ANY ban, as these procedures are more statistically proven than our current methods and base their results moreso on facts more than the current procedures.
So are there any thoughts/ideas about this idea? I'd prefer if people with more knowledge towards the field of statistics would make themselves known, as they can probably analyze this process best (and probably edit it better than I can, as my statistical experience is limited).
Because there has consistently been talk of having MK banned, if anyone wants him banned they should find numerical and clear data based on facts instead of unbiased claims. In order to best do this, we cannot reject the MK = not broken hypothesis until we have adequate data to prove that he is broken, or that the metagame revolves around him.
First, we need to determine what constitutes the metagame revolving around him, and I think its fair to say that it's generally seen that the characters who do well against metaknight are higher in the tier list and in general get better tournament results rather than most characters who don't do well against him (ie Snake, Diddy, Falco doing better than ROB, Marth, and Olimar). However, this does not disprove the current hypothesis that MK is not enough of a central metagame unit to warrant a ban. I AM NOT SAYING THAT MARTH ROB AND OLIMAR ARE ON THE SAME LEVEL AS SNAKE DIDDY AND FALCO AND I ACKNOWLEDGE THAT THERE ARE OTHER FACTORS THAT GO INTO TIER LISTS.
So, what we want to do is compare the tournament results of a set of MK banned tournaments to the standard tournament results that we have collected recently (as to accommodate for a modern metagame). We should compare the rankings of characters in tournaments without MK to rankings of characters in regular tournaments, and use a Z-Test to see if there is any statistical significance in a character's placing between MK banned and MK legal tournaments (naturally, this will take several months of tournament testing in order to gain adequate data, and hopefully several people with experience in Statistics to collaborate with the Z-tests, determining what is statistically significant/choosing an appropriate alpha-level, sampling, checking work, etc.)
Now once we determine WHO is affected by this test, we need to find out if those characters being affected is enough to warrant the ban. Say only Ganon benefits from MK getting banned, or, in a less radical example, Ganon, CF, Yoshi, and other characters that, even with MK banned, are hardly tournament viable. Is that enough to warrant a ban? Of course not. What about Lucarios, ROBs, Marths, Olimars, and Toon Links? Is 6,7, or 8 characters getting significantly better, or even worse, worth the ban? A limit needs to be determined, say n number of characters, of characters that have statistically significant differences in placings and usage in MK banned tournaments, in order to warrant the Metagame revolving around him.
Ideally, we could determine this n value comparing it to tournaments where a character of similar nature (Snake, Marth, DDD, Diddy, etc.) is banned and testing how many characters are significantly affected by that character's ban, but that occurring is very unlikely, so I hope that should MK banned tests like these actually take place, the back-room determines a fair n value instead of running multiple character banned tournaments (realizing that as less tests take place, much less/more characters will be affected than the true result).
Perhaps even two levels for n characters could be obtained, one for determining whether the metagame revolves around him, and one for determining whether the metagame revolves around him too much.
Naturally, there are multiple ways to do things and I am sure everyone is open for slight modifications of this process, but I suggest we go through procedures like this for determining ANY ban, as these procedures are more statistically proven than our current methods and base their results moreso on facts more than the current procedures.
So are there any thoughts/ideas about this idea? I'd prefer if people with more knowledge towards the field of statistics would make themselves known, as they can probably analyze this process best (and probably edit it better than I can, as my statistical experience is limited).
