Here is the problem, and it is elemental to the way that Steem/Hive works:

It is through the process of making and taking the bets that the value of a post is determined. Coupled with the proof of stake system, that effectively means that the same people with the most power to affect the payoff are the ones who have the capability of maximizing their payoff.

The problems with that are obvious and you almost/sort of touch on that in your very last paragraph as almost a throw away. It's a central issue, however. It is the crux of the whole problem. Curation and reward have been inextricably bound in the underlying system and to fix it, that is to remove behaviors which result in undesired results, the whole of the system has to be modified.

Unfortunately, that can't be done at a mechanical level and still be able to refer to the result as Steem/Hive (and I keep referring to the systems as they stand like that because they are identical in implementation). It is certainly possible to imagine a system which decouples curation and tipping/reward and it's probably a good idea to do so because it would have a number of positive knock on effects.

I just don't think it can be done here.

Specifically as regards downvotes, the idea of "free downvotes" as it stands already violates one of the central axioms of the system, effectively that your stake determines your ability to affect the distribution of the rewards pool. Giving people free 100% downvotes means that your stake is effectively multiplied when it comes to reducing the power of other people to be rewarded from the pool. If it were simultaneously matched by the same number of free upvotes, at least there would be some sense of parity.

At that point we are just back to giving every new account 15 SP for free, which at least is a lot easier to keep up with and see the immediate problems with, as opposed to providing a multiplier for SP which is semi-obscure.