Right now? the standard deviation of all of the data points is 12. So I had it set to 12 points separating the tiers. 0-11 for B tier, etc.
For example,
Event Hubs tier list (which is an accurate culmination of players from all skill levels using various different experiences to aggregate) currently has an SD of 14. This being applied to their list would be the following:
A - Gen Fu (+25.1), Ayane (+24.1), Christie (+21.2), Sarah (+19.1), Kasumi (+17), Helena (+14.5)
B - Leifang (+11.3), Pai (+7), Jacky (+6.8), Jann Lee (+5.3), Momiji (+5.1), Rig (+4.7), Hayate (+4.5), Tina (+2.9), Hitomi (+2.7), Alpha 152 (+2.6), Mila (+2.6), Hayabusa (+2.3)
C - Phase 4 (-3.0), Akira (-3.6), Zack (-5.1), Lisa (-5.4), Marie Rose (-7.2), Rachel (-11.8), Kokoro (-13.2), Brad (-13.8)
D - Bass (-14.1), Nyotengu (-14.2), Bayman (-14.9), Leon (-16.0), Ein (-18)
E - Eliot (-38.4)
In the end it’s typically subjective on how the characters are sorted in tiers. I did a lot of research looking into how match up tiers were reported and couldn’t find a consistent reasoning for division besides just what looked like “Yeah this is where the line would be divided”. So I opted for a more scientific method by finding the standard deviation amount from the data and then using that number to divide the groups.
Also, the FSD tier list chart has already had some match-ups altered, so of course some of the characters have been adjusted in the tier list. But the FSD tier list only has 1/3rd of the character roster submitted and discussed by some of the top competitive players in the community. Another 1/3rd are still waiting to be submitted and then discussed when conflicts arise from other submissions. And the last 1/3rd either don’t have a competitive player representative of them or will just fall in line based on the submissions of the other 2/3rds.