Good question! Unfortunately, there is no 100% consensus surrounding what the policy should be. Some advocate that of course we should encourage useful edits, others argue that we should discourage people from making many edits. I'll just give my personal opinion. Everything from here on is my view only.
tl;dr: I think moderators should grant all requests to remove auto-c.w. status as long as there is no pattern of abuse (e.g., too many trivial/insignificant edits).
Encourage substantive edits. My belief is that of course we should encourage all substantive edits that improve the post. Part of encouraging such edits is to ensure that moderators "have your back": if a sequence of such edits triggers auto-conversion to community wiki, then in my view, of course moderators should be happy to remove the c.w. status (just flag it for them). That's what I'd like to see happen, anyway.
Justification: our mission is to create a repository of high-quality questions and answers that will be useful to many others. Therefore, edits that improve the quality of questions or answers serve that mission and should be encouraged.
Avoid trivial edits. However, one thing everyone seems to agree on is that you should avoid making edits for the purpose of "bumping" your post to the front page. That is universally viewed as abusive. If you ever find yourself tempted to make a tiny edit to get your post more attention or to push it back onto the front page, you know you've gone overboard -- absolutely do not do that.
Of course, this is not directly measurable for anyone else. We can't read your mind and divine what your purpose was. Therefore, we have to assess your edits based upon their content. For this reason, it is best to err on the side of making significant, substantial changes to your post -- that way, no one can argue that you made a tiny tweak just to bump the post.
In my view, if you're making substantive changes that improve the post -- even if it is something as minor as adding an additional citation to the literature, clarifying the post in response to a comment, adding an extension you thought up on the walk home, or editing to clarify an imprecision you noticed on re-reading your post -- then moderators ought to grant all requests to remove auto-c.w. But if you're repeatedly making minor changes -- e.g., fixing an inconsequential spelling error, making minor changes to the formatting, adding bold tags, etc. -- then you are now in the grey zone and moderators might reasonably deny such requests.
Make the first draft count. As a general practice, I also encourage you to make a reasonable effort to make the first draft be as good as you can make it. Proof-read your question/answer using the preview pane before posting it.
You don't have to go overboard. You don't need to draft it offline and sit on it for a day (or re-read it obsessively) before posting. We don't want to discourage you from posting an answer.
But at the same time, don't use the ability to edit posts as an excuse for posting a useless first draft. On high-traffic sites like Stack Overflow, I've seen some people first post a one-sentence answer, then repeatedly make many small edits to add to their post, in hopes of getting the first answer in and getting free upvotes that way (the "Fastest Gun in the West" phenomenom). I rarely see that here, so I probably don't need to mention it -- but I think it is generally best to avoid that practice. If you do do that, don't be surprised if the moderators are less open to requests to remove c.w. status.
Requiring people to edit offline. I would not suggest we ask people to draft their answer offline, as a condition of removing auto-c.w. status. I don't think it's our place to dictate workflow to authors. Some might be happy to draft their answers offline, but others might find that less attractive. I wouldn't want to drive away people from contributing good content -- I don't think that's something we as a community should be doing, and I don't think we should hold the "c.w.-removal" hostage to such conditions.
I think it's perfectly fine to mention to people that, if they want to avoid auto-conversion to c.w., one approach is to draft answers offline and wait to make edits so they can batch them up -- but I don't think we should make this a condition of "c.w.-removal".
I understand this is consistent with how our moderators are already handling things, which is great.
Your specific case. Your answer is a textbook case of a reason why moderators should be delighted to remove auto-c.w. status if it gets triggered by one of your edits. Your answer is fantastic, detailed, a real gem. I wish I could upvote it a dozen times. It sets an absolutely fantastic atmosphere for this site -- I wish we could have more answers that were one-tenth as comprehensive and helpful as your answer.
Yes, you made many small edits to your answer. However, each edit that I looked at was substantive and did improve the answer -- and any answer of this extreme length creates more opportunities for improvement and thus is more likely to run into the threshold for c.w.-conversion. If editing a figure triggered auto-conversion to c.w., all you'd need to do is flag it for moderator attention and ask them to undo the c.w. status -- I would certainly hope that moderators would gladly grant your request. (If such a request was denied, I would be up in arms and extremely upset!) It sounds like the moderators are happy to do this.