TL;DR: Drasticness cannot be the sole disqualifier of an edit. Make edits that you think improve an answer's correctness and completeness, even if drastic. Expect some to get re-edited or rolled back. Assume good faith and don't get your nose out of joint.
As the perpetrator of both and a relatively new member, I hope I've got some valuable perspective to add:
I see the two examples as representative of two categories of large1 edits: explaining and completing2. Both strive toward what I think is this site's hallmark: providing correct and complete answers to well-posed questions.
Explaining
Short answers obviously can be correct. And, as this is a site striving to be an authoritative source for correct answers to well-posed questions, for many visitors a one-liner is perfectly-suited to their needs: "if it's been up-voted on SE, that's enough for me!"
But some questions--perhaps most?--are requiring of explanation and references to common material. For the user who comes not with the question "can a monk..." but with the real question "how do I know the Monk can target two..." or "where can I point a monk 2/rules-lawer 30 to do his own research first", the explanation is the completion.
Clearly I thought the first example could use some explaining, as the sticky point for someone in doubt--in my mind--was "when the monk makes the bonus attack is it some continuation of the previous attack, or a brand new attack action in its own right?"
But I've violated what probably should be a style guideline: answer first, explain second. The way I've edited it the argument builds to the answer, rather than answer-with-supporting-explanation.3
Completing
In the second case the question specifically asked for variance among editions. The accepted answer compared two versions. Most other answers each addressed one, and none addressed the 5e situation. So should 5e information spawn a new answer, or be folded in to the accepted answer? Even recognizing that addition of 5e increased the accepted answer's length by 50%, it seems to me that 'accepted answer with three versions' is better than 'accepted with two versions, filter through the other six4 answers to find any other versions you're looking for.'
From there it's an obvious leap to pull in SSD's 2e information and a distilled version--in keeping with the answer's style--of the 3.5 info. (Again: I botched that. Mea maxima culpa.) But assume I had done it correctly: we'd have an accepted answer with editions 2-5 represented, at triple the length of the previous revision. Is that better or worse?
Here's how I see the situation: there's a lot of different, correct, useful information dispersed among many answers: pull it all together. perhaps this is the situation where we need a community wiki, rather than 'an answer?' I don't know enough about it to know the propriety, or even how it gets done.5
So, 'When to Edit?'
The edit help itself, particularly at "When should I edit posts?" strikes me as... conflicting, at best.
Let's keep in mind the example of adding 5e info to the second question: is that edit "substantial and [leaving] the post better than you found it," or does that edit run afoul of the bullet points, all of which are much-more minor that the edit I made6?
Two frameworks that have guided some of my thinking are WP:BOLD and WP:OWNER. I know we're not Wikipedia. But we're a knowledge and information site, and ignoring the precedents and culture of the English-speaking world's dominant information site can only be harmful.7
Factors in Support of: Be Bold, No Ownership
- the edit button prominently at the bottom of each post
- the assurance that edits from new users will be reviewed, encouraging even the least-reputable among us to attempt edits
- the existence of a revision history guarantees edits can be thoroughly reviewed, including to the level of moderator roll-back
- edits are noted in the by-line, encouraging readers to look at revision history
- edits move a post up in the "active" listing, further encouraging review of the edit
- original authors may not be responsive, or even active
Factors Opposing: Be Bold, No Ownership
- the original author receives permanent by-line
- the original author receives all reputation that will ever accrue to a post
- WP's 'no ownership' policy usually comes into play not in situations like this, but with a much-later user deciding "I'll assume the (sole) responsibility for curating the edits to this article." (Although maybe it appears that's what I did... I didn't intend that.)
- a highly reputable user--miniman--explicitly: "with drastic edits like the ones you made to this answer you should ask the person who wrote it first." (8/19/15 1511 comment on second example)
Conclusion
Despite the opposition factors, I think the safeguards and site infrastructure argue for a permissive attitude toward even large edits. I don't want to end up with 'edit-wars,' but I think our ideal should be correct and complete.
Footnotes
I say 'large' rather than 'drastic' as I think the question of "what makes an edit drastic is a matter of personal judgment. As evidenced by the various approvals/rejections/rollbacks of the two in question.
Obviously, the second example is badly polluted by the gross error I made in reading the 3.5 situation. But I think it's clear--especially looking at the revision history linked--that my intent was to flesh out the accepted answer.
In part the structure flowed from a desire not to change the ordering of the original material, in part it's just my poor style.
As adding a 5e answer would have brought the total to seven.
Though if you, Dear Reader, do think that's the solution please edit the answer's line to make 'wiki' a suggestion rather than question and remove this footnote! Improve the answer!
To be sure, one bullet point mentions making updates, but prefaced by "minor mistakes"--certainly the feel I get from the bullet points is "keep it small, buddy!"
I'm not saying we need adopt any of that site's policies, but we should absolutely be informed by them.