You are viewing a single comment's thread from:
RE: Towards a decentralized, abuse resistance framework for the Steem blockchain
I think this idea's brilliant and in many ways, builds upon the work that the STEEM Watchers Team is already doing.
I've got a few additional thoughts which you may have covered so apologies if I repeat any:
- Anonymity - I think that this is important. Whilst the game needs to know who's playing, the community doesn't. This aligns with the quorum sensing (I'm surprised you waited until so late in your post to mention this 😉) which tries to protect smaller users.
- Player Scores - To avoid "abuse" of the game, if a user's opinion consistently differs from the majority, then they should have a "Rating" where their opinion carries less "weight" than a "sheep". Working along the lines of Steemit's "Reputation".
- Let everybody play - It will be a couple of years ago now that I suggested writing an "Anti-plagiarism" game, where one player would submit a post for review (with evidence) and then other players would simply pick "plagiarised" or "not plagiarised" - the idea being that "Watchers" could the flag the most "plagiarised". And to gamify it (with an anonymous reward system) would attract players. I wrote my alternative front-end instead 🙂
I'll stop my thoughts for now... my brain's consumed with the plethora of stylesheets that I'm discovering in GitHub repositories that aren't Condenser!
Yeah, it didn't occur to me until that section of the article, but this really is just another proposal to implement quorum sensing. The surveyors are sending the signals, but nothing gets acted upon unless their signal is strong enough to find its way through the analyst and enforcer filters. I definitely agree that anonymity (or at least strong pseudonymity) is important, at least for the surveyors. Pseudonymous or not, protecting the analysts from retaliation would probably fall to the enforcers. That's true with the current STEEM Watchers team, too.
You're right that STEEM Watchers has a similar model already, where the detectives are basically acting as surveyor + analyst. The concern I have with their model is that it creates an incentive for a Steem Detective to create abuse with a hidden account and "discover" it with their detective account. I'm not saying that anyone is currently doing that, but the incentive is there.
Yep. I remember that. I think that conversation is probably what started my thinking about the swipe-left/swipe-right model. I just think it needs to extend beyond plagiarism.
Yeah, a lot of thought needs to go into this, and also into the beneficiary reward distribution. I'm also not sure if someone could skew the results by running multiple surveyor accounts. Hopefully, the random assignment of posts to surveyors would prevent that.
Glad to see that your proposal got funded. Congratulations! I look forward to reading about your progress (and seeing it implemented)!