If, for whatever reason 'good' or 'bad', they happen not to believe that @themarkymark should receive rewards, it's pretty unlikely that @themarkymark receiving rewards could ever be said to have consensus. You can adjust things around the edges somewhat, but that fundamental fact is hard to avoid.

True. If I adamantly assert that a given post deserves high rewards and you adamantly assert that it deserves zero, then there can NEVER be a true consensus (according to the definition I provided above) on that given post, because I will never 'accept' your desired outcome and you will never accept my desired outcome.

The challenge is, can we adjust the mechanisms to allow 'disputes' such as the one between marky and newsflash/xeldal to be determined more widely rather than just mano-a-mano.

The current system affords supporters of marky only two options: [1] upvote marky's posts and consistently lose curation rewards or [2] employ downvotes against other posts, with the express aim of inflicting curation-rewards damage to the original downvoter(s), to try to dissuade them from persisting in their attacks on marky's posts.

A Counter-DV option (like the one originally suggested by @theycallmedan in the aforementioned post) would allow others to costlessly pile on Counter-DVs (in accordance with and limited by their respective stakes). This might then invite others to pile on DVs. In the end, we get a truly stake-weighted near-consensus on those 'contentious' posts.

My suggestion with respect to Counter-DVs would be to have them either be weaker than DVs and/or less frequent. As such, the total 'value' of DVs able to be implemented would be greater than the total value of Counter-DVs. This allows DVs to be the dominant mechanism for combatting online abuse. However, it provides a mechanism for disagreement over rewards to be more balanced, because stakeholders can costlessly throw their stake into the decision-making process, both for and against rewards going to a single post.

For what it is worth, number of accounts does not matter at all. A swarm of 10000 low stake bot accounts should have little to no influence. I prefer to speak of stake or some other demonstrably effective reputation mechanism (extremely hard), and not 'accounts'.

Agreed. When I said 'accounts' I was referring to real people with real skin in the game. And stake-weighted skin in the game is as good a 'reputation' metric as any, imho.

The problem with the current system is that those voting 'for' a contentious post (via upvotes) do so at their own loss, whereas those voting 'against' (via DVs) do so at no loss to themselves and force a loss upon their enemies.

We need some balance in this regard.

If whales upvote and bestow a large windfall on a post (in fact even a single whale can), then it's fitting that other whales can also downvote that same post, no?

Yes, that is fitting, provided there is some community-wide balancing mechanism, rather than a mere unilateral downvote.

Let's consider two extreme cases, under the scenario where Counter-DVs are free and are worth 50% of a full DV, but no self-supporting Counter-DVs are allowed.

First, Alice publishes a one-sentence post saying "Good Morning!" Bob (a whale who knows Alice or maybe is Alice) upvotes the post with his $100 full upvote. Charlie sees this as entirely inappropriate and issues a -$100 full downvote, fully negating Bob's upvote. Bob cannot reverse Charlie's DV, and no one else in the community is willing to support Bob's clearly inappropriate behavior.

Second, Alice publishes a well-written post. Bob (a whale who knows Alice or maybe is Alice) upvotes the post with his $100 full upvote. Charlie sees this as entirely inappropriate and issues a -$100 full downvote, fully negating Bob's upvote. Bob cannot reverse Charlie's DV, but asks for help from the community. Ten accounts see Bob's request but agree with Charlie and pile on an additional -$50 worth of DVs. Twenty accounts agree with Bob and pile on $100 worth of Counter-DVs. The net result is that Charlie's attempt to zero out Bob's upvote on Alice's post was partially effective. Instead of the post receiving $100, it only received $50. However, it took $300 worth of voting stake (Bob's $100 upvote plus the twenty accounts giving $100 in Counter-DVs, which required $200 in combined upvote power) to award Alice $25 in author rewards and Bob $25 in curation rewards (which represent 50% of what Bob would've received had he upvoted a non-contentious post).

Although the second scenario does not technically meet the definition of 'consensus' (because Charlie's plan to zero the post, which he was adamant about, failed, as did Bob's plan to award $50 to Alice and $50 to himself), it is much closer to a 'consensus decision' because it represented the stake-weighted combined actions of 32 accountholders instead of just 2.

To me, this comes much closer to 'consensus decision-making' even though it does not strictly meet the definition of consensus.