I view deletion as a last ditch effort. While I did not evaluate the thread, because I found my ability to interpret the reliability of scans to be essentially next to nil, I do recall the thread Dark Soul is talking about aiming to rectify the situation outside of just outright deletion. Broadly speaking, I would find that sort of solution to be better than applying very wide deletions of content. I rather like it when people's work gets to stay up.
That said, it is a legitimate concern in determining what exactly is MTL-spawned and what is not, and so far, much of the discourse on that has amounted to personal memory rather than anything concrete. I'd like to hear from someone knowledgeable whether it's fairly easy to determine that sort of thing or not, since it seems if we were to avoid deletion, we would need to be fairly confident in the status of a great many scans.
As for your actual proposal: the above speaks to my thoughts on the deletion. I would be content with requiring scans to be passed through our translation helpers (although one fears we shall run dry of reliable translation helpers willing to sift through enormous amounts of scans; they may find themselves overwhelmed, and it seems actual Japanese speakers are fairly rare here). I'd prefer to hear the thoughts from a Japanese speaker on whether these mistranslations you bring to light would be reasonably interpreted as a mistake: if you speak Japanese yourself, Reiner, I would happily take your word for it. Otherwise, I would like to hear from someone who does, and only then would I approve of doubling the duration of the ban (or applying some other extension aside from your recommendation).
EDIT:
I do see that the thread Dark Soul is referencing has made very little progress. Still, if it could be finished to satisfaction, I'd prefer that to deletion.