I liked the suggestion you made to my original idea on this point. "I think it would be possible to take the downvote curation reward penalty only from those upvotes chronologically before the downvote, so upvotes to counter the downvote wouldn't be penalized."
Not sure exactly how we get there, but my point remains. If you get diluted by taking action, very few, if any, will take that action. It helped a lot with DVs. While I agree we need more healthy DVs on the most post, it's rare to get a post on web2 without a downvote. It's just people need to adjust the downvote size, or it causes harm. No need to zero out a post from someone legit every time they post to the chain. So it would be nice to find a way not to be diluted when trying to find equilibrium with the token distribution.
I saw a comment here I forgot who from about a possible inflation reward pool for DVs. Meaning if you don't use the DVs, you miss out on rewards. I believe nudging good people to use downvotes will be very healthy. It's ok to do a small/normal size downvote in relation to the post rewards if you disagree. And if we had a more balanced healthy DV system, we would see much more competition for the rewards. Obv this would also help "bad" downvoters, so my point above remains. Maybe a mix of both, one to encourage healthy DVs and one to help counter overhanded ones.
Downvotes spread more widely would indeed help with some cases, since then getting downvotes would not be so devastating to rewards (most content would have some so the zero-sum aspect would mean it would tend to balance out some).
Though it still wouldn't help with these "downvoted to zero" situations. Again, people have to accept that in a voted-reward system, highly controversial content with a lot of disagreement won't do well on rewards. We don't (for the most part, though front ends do hard block child porn, etc.) censor it, but we won't reward it either.
Layer two where a narrower subcommunity (i.e. with less disagreement on some things) can decide differently about what it wants to reward is probably the best solution.
Imagine an economy where you had the option to EITHER give performers a tip OR take a tip from a performer.
Now imagine that when you take a tip from a performer, they're tip gets "redistributed" to the remaining performers, with the largest share going to the performers who earn the most tips.
Small performers will be heckled because there is no fear of retaliation.
Large performers will always reap the majority of the rewards kicked out of the hats of the smaller performers.
In the end, you're simply reinforcing an anti-competitive environment where the big fish use their position to only allow performances they personally approve of.
Sure, there may be some legitimately "bad actors" - but even they deserve some pretense of a FAIR public trial and some reasonable path to redemption.
Secret DISCORD kangaroo courts do not meet this standard.