From our tests and research so far your system will probably work until it doesn't. I.e, there will always be malicious actors looking to break it without the correct incentives not to. The biggest danger would be external griefing attacks from those that don't care about your platform. Assuming the token or stake is tradable, it's likely a large enough quantity could be bought that you have some sort of sybil or 51% attack on the voting system.
Ok so, the tokens (ZCH) you are able to pre-purchase instead of paying the posting fee while saving a new message to the ledger are freely tradeable. However, those tokens don't add to your overall voting rights. I'm sure with enough evil intend there will be people coming up with the most entertaining ways to try and break the voting system. So your research was as good as ours; it works until it doesn't. But as you can see in the whitepaper we reserve the ability to change the entire moderation contract so we can update it or completely change it with new routines as time goes on. We're currently figuring out a good way for the users to be notified in advance if a change is upcoming and how to integrate them into the process of voting the change in or out.
There are other projects out there other than us of course, so the final answer may be elsewhere for you guys. Assuming we attain a large enough juror pool and dispersion of tokens, it makes it harder to break. I guess if you manage to disperse the tokens in a nice manner this will mitigate but the griefing of an actor with reasonable funds is hard to stop. Our mechanism against this is forking but I'm not sure if you've considered that.
The point of 0xchan is not to amass riches. The dividends functionality exists to basically repay people their posting fees over time. I'll crunch some numbers during the weekend but from what I have in mind your spending/earning ratio is so out of whack that it makes no sense to collect massive amounts of ZCH just to gain dividends.
The cryptoeconomic game: It's a tough one.
It is. But it's also a fun and interesting one. In 2018 alone more economic models have been tested in record time thanks to various blockchains, than we could ever try in "real world" economics. Gotta appreciate that.