I think it's a very difficult thing to put a percentage on. We don't even know here whether someone has actually successfully CTB using it. Just because there is a goodbye thread, then nothing- doesn't necessarily mean they were successful. They may have eventually used a different method even. I guess there are some 'official' case studies but probably not many. It's not something any official organisation is going to want to release data on I imagine.
I've heard people have vomited- even after taking anti-emetics. I've heard people say it doesn't necessarily mean the attempt has failed. I don't know personally.
I would say- people who come up with percentages ARE very likely consisering all the cases where people were either found, or called for help themselves and the process was reversed via medical intervention. That I find really bizare. It clearly isn't a fair way to assess the effectiveness of a method. Many methods would fail if the person was stopped part way through! It doesn't necessarily mean that the method is unreliable. Seems to me like the equivilant of saying that all people who are stopped from jumping made a successful attempt but that simply isn't the case. We can't judge whether the attempt would have killed them because they didn't get to do it!
From what I've seen here- failures tend to happen when people don't follow the protocol. Or- it wasn't SN they had. I just don't think it's that realistic to try and put a percentage on it though. Not when so much of the 'data' is supposition.