I guess by perfect I mean whatever you consider 0%, the question being that should we expect to have to only clean, or should we expect there to bubbling on some models and we just have to deal with it either with green stuff or replacing it, and if so what is an acceptable percentage.
For now the sample space is too small to use this as evidence but if we could get a few hundred the importance of the poll in giving a rough idea of how many models people are finding defective increases and if we could get a thousand plus votes it would act as a very good piece of evidence.