Chou Toshio said:Since there's nothing measurable, and we're just finding about opinions based on current metagame conditions, there's nothing wrong with testing multiple Pokemon at the same time-- it doesn't create error, because other "suspects" are also a "metagame condition."
Well, in fact there is an obvious bias in multiple pokemon test. Take Torn-T, testing both him and Keldy in the same time. Testing multiple suspect at the start of metagame where there was Darkrai and other his one thing, but here it was another.
People found Torn-T the more threatening (no problem here), and then they focused on him. Keldeo was considered a little bit secondary. The comparison between the two pushed people to vote easier ban for Torn-T and easier OU for Keldeo. Also, Keldeo's result was his result in a Torn-T metagame. As soon as he got banned, the fact that Keldeo was deemed OU lost a little bit of his value since he disliked Torn-T. Ultimately, I don't think this bias changed the result of bans, but surely it had an influence on some individual votes.
And if such bias are allowed, they'll end to, in some test, to change the result.
Talking about statistic is maybe not the question here, but bias IS a concern, even more when it's a subjective choice. (Also I'm just saying bias are a real problem which shouldn't be overlooked, not anything else)
tl;dr : each way suspects are done lead to bias, but not at the same point. Bias is
never good. Multiple suspecting is problematic, polarizing opinions.
Agree with ginga's analyse which show good Keldeo's recent history in OU.