I have really only skimmed this thread, but if you are deciding whether to reset ratings or to have a seperate ladder, then I made a suggestion in my stage three and beyond thread that might be useful here..
What we need to do is during this period calculate two seperate ratings for battlers on the OU ladder. The normal rating, which wont be affected at all. And a second "suspect rating" which would be reset to the default position at the beginning of each test. This way people wouldnt complain about losing their ratings, but people also wouldnt qualify without a lot of battling during the month.
[edit 2] - Just to make sure this is clear, my normal rating would continue unaffected, but at the start of the month I would be given an additional rating, that would start at the default position, and change with every battle as though I were a battler with rating equal to that suspect rating facing another battler with rating equal to his suspect rating.
Maybe this would require a lot of work to implement? I have no idea, but it seems to me to be a perfect solution otherwise. Even if it was a lot of work in the short term, I think it would be worth it for all future tests on any ladder..
[edit] - it seems this thread is actually discussing exactly what my thread was intended to discuss so perhaps I will read things further.
Yeah, I'd like to argue in favour of starting with suspects legal and banning them one by one.
Practically it is much easier, you start with the current ruleset. Which I think is a huge advantage, especially if you are testing on the main UU ladder, it is less disruptive to people who are not interested in the testing process; but also it means less work for the battlers who have to deal with a less violent change at the start of the process.
It also forces people to participate in the test, for instance you cant refuse to use the suspect and quit battles against people who do.
It gives you a better idea of what the suspects are capable of, since they are being used throughout the testing period, hopefully they will be tested more thoroughly. If you start by voting on the most obviously uber pokemon, then the more difficult decisions will have had much more testing than they would otherwise.
I think that in the case of OU testing the stage 3 test has been quite a sudden shock to the system, I mean to end the testing process with a test that seems likely to be different from the final ruleset seems backwards, just intuitively it seems more logical to start from where you are, and then work systematically towards your final goal.
Have a nice day.