This test isnt about what people want, or else you would never have had to have polled anyones reasoning. In this context people are using the word broken to mean "far too powerful to be in line with a balanced metagame" which is exactly what this suspect test is about.
I was using Tangerine's words there, obviously. What people "want" should be in line with our philosophy, or else we could just have people voting to keep things out of standard because they aren't easy to get ingame or something. And such anomalies are addressed by the fact that we...
I am a qualified player, and if you were to poll only me we would have an extremely biased majority that nobody should accept. The point here is that any poll is imperfect (some moreso than others). But that it is better to err on the side of more testing when there is a reasonable likelihood that the vote gave a false result (based on our definition of uber).
...ask the opinion of more than one person. Even if I could chastise you openly for not being able to comprehend our Characteristics of Uber that have been stickied at the top of this forum all year long and that you are in de facto agreement with by playing on the Smogon server, it wouldn't really be nice for or fair of me to do that if your singular vote didn't make up, say, 3% of the total pool, right? (Hypothetically, of course.)
We keep testing until no suspects are voted uber.
If a pokemon cant get two simple majorities in a row on this test then it clearly isnt "very apparent" enough that it is "far too powerful to be in line with a balanced metagame".
The argument for a supermajority is essentially the same except it is slightly less clear that it isnt "very apparent" enough, but obviously still clear enough for a lot of people (this isnt the course of action I support here).
What I was getting at is that it seems possible for testing to stretch to Stages 3-3 and 3-4 on the same "statistical anomaly" deal that upset Garchomp's simple majority in the first place (which would be for the same reason these are Suspects in the first place), but I'm prepared to let things play out and hope I'm wrong.
But manaphy will be retested and latios was voted uber by more than 9 votes. The point of the supermajority is to protect from this sort of thing, to increase the likelihood (to a point where there are reasonable doubts as to its legitimacy) that, even though the poll (like all polls) is imperfect, the pokemon will only be considered uber if it meets our definition of uber.
Have a nice day.
I actually meant Skymin not Manaphy (which is clear when you take it from the context of the paragraph above, both pokemon were voted uber in Stage 3-1). Regardless, things become shady when we don't know what "supermajority" we are striving towards until after the fact, which is the reason I kept posting in this thread to make sure this was addressed before Stage 3-2 concluded and not just a vague "later".
Anyway, after a chat on IRC, we've decided to test Garchomp in Stage 3-2 as soon as Doug is available and can restart the server. I look forward to the completion of another round, and as always, implore the rest of you who have posting privileges in this forum to use them as soon as you deem it necessary to avoid unnecessary delays to changes on pokémon policy.