Thanks for that.Supermajorities are important because they indicate that the result isn't a swing result. Say you have 99 total votes and something receives 50 votes to in favour of a ban. A single voter swing (which could've been achieved if the vote had e.g. been held a week earlier or later or if one person hadn't slept through their alarm clock on the last day of laddering or whatever) would have been enough to change the result from a majority to a minority, and if there had been one extra voter it could have just as easily been a 50:50 split. As such, it is safe to say that taking action would be controversial and statistically insignificant. As such, the burden is placed on the party looking for change to demonstrate that it is decidedly the most desirable option—or, more accurately, the option that most people say they want, as what people say they want does not necessarily align with what they actually want or what is the actual correct course of action; gamers are great at spotting symptoms but not very good at identifying problems (case study: Jago in Killer Instinct (1:19:17–1:21:40; timestamp linked)).
A supermajority is an arbitrary point at which we deem a majority to be statistically significant. This could be 11:9 (11/20 (55%)), 3:2 (3/5 (60%)), 2:1 (2/3 (66.67%)), 3:1 (3/4 (75%)) or whatever other ratio, but as a broad rule of thumb, a higher supermajority requirement increases the significance of a vote in exchange for making it harder to enact action. IIRC the one used by OU is 3:2; my personal preference is 2:1; depending on who you ask, you will get different numbers for where they say they believe the sweet spot is, but whatever is set for a specific suspect test is all that actually matters. Smogon doesn't have the resources or even necessarily anyone qualified enough to do anything as in-depth as a paid game developer working full-time in the balance department, but we can at least get a loose image and conservatively trim the edges to get something that is okay enough provided we don't have too itchy of a trigger finger, which is what supermajorities aim to inhibit. Of course, for an environment unlike OU, a simple majority may be preferable, either because the format is highly unstable (e.g. lower tiers with frequent tier shifts) or because the playerbase is small to the point that a lot of supermajority benchmarks would end up being swing votes regardless (e.g. most OMs).
I personally think supermajority would be much more needed in lower tiers than for OU because there is way less voters for lower tiers than for OU suspect which means the chance of getting 50% on both sides is far greater than for OU. For this generation, almost all OU suspects reach at least 100 voters (the least being Roaring Moon with 97), this is something way above lower tiers.
If we look at last suspect for each lower tier suspect test this is what we have :
SV UU - 44 voters
SV RU - 21 voters
SV NU - 20 voters
SV PU - 25 voters
I don't know what is the best, but I feel like 60% or 66.66% (basically 2/3) is way too high when you have that many people who're voting (talking about OU here).