1- While this train of thought is right, it's only true when your whole community adapts at lightning speed. If it's possible to abuse, then 1 person discovers it, then the next day everyone does it, then the token value instantly crashes to $0.
In reality, adaptation will take time, and the early malicious users would profit at the start of the crash, while the people who acted good would just lose everything because they acted good. It's a bit similar to the problem of bid bots on steem. If there was no bid bot today, someone would make one because an opportunity exists, and as long as not everyone sells his vote, the token keeps on existing and being worth a little something, and the thieves keep on profitting, even if it's little, they don't care if it's automated.
Overall I think reputation is a complex problem and I think for a website like yours, trying something new wouldn't be a bad thing :D
2- The way you calculate the ReviewScore seems a bit prone to sybil attacks but considering theres only a *2 factor for the VotePoints part (which is the part sensible to multi-accounts spamming), it should be fine.
I'm not sure I understand, taking the average seems a bit silly. Shouldnt the overall score of a project be the sum of the review scores instead? I mean, if a little minnow creates a new review and upvotes himself for 0.001$, it will lower the average and bring the project down in the rankings?
3- I think it depends on what you want. I don't think curation is viable on steem at 25%, so at 20% it won't be. So in this case you should go for 100% authors. You'd get tons of reviews but then there would be no incentive for people to vote on reviews which kinda defeats the purpose. Anyone can go on your site and create a review and earn from it, without even putting a cent, but in order to make money with curation you need to invest, so I think they need to have a better incentivization than 25%. I'm not even sure 50% is enough, I run an experiment on DTube where we redistribute our beneficiary rewards to increase curation rewards, and while it has a bit of effect, I don't think it's enough.
4- In my view, doing an ICO is just a way for you to give a discount on your tokens. If you believe in your token and it's future utility and use, then it's probably better to do like Steem did. Keep your tokens, then sell a % of it every year to fund the development and so on.
I'm glad to see a real website working on SMT implementation, instead of new steem projects popping, it's a breath of fresh air :D
1 - Got it! Are there any other rep systems aside from Steem-UA that you'd recommend we investigate?
2 - Agreed on Sybil attacks; we're trying to strike a balance. If we don't give enough weight to up-votes in the
ReviewScore, then an honest minnow's up-vote would have 0 impact. So we figured we would give a 2x factor onVotePointsand see how it behaves in production. If we notice it's not working, we can always adjust the algo later and re-calculate because it's all run and computed off-chain.Check out the examples on page 17 and 18. In traditional review systems, an entity that gets a 5-star review and a 1-star review would have a net overall rating of 3 stars because there is no differentiation between each review, they are treated equal.
(1 + 5)/2 = 3. We attempt to differentiate between high and low-quality reviews via theReviewScore, so a crypto project that gets a 1-star review and a 5-star review is likely to not have an overall score of 3, as one review may carry more weight and therefore skew the overall rating in its favor.3 - Got it. I'll do more research on the author / curator split. Are there any other steem-based dapps that you know of that are struggling with the 75/25 ratio?
4 - Our view on the ICO is mixed. We'd like to inject capital into the project ASAP to accelerate development, but we also fear selling 55% of the token supply to a small pool of investors who's primary motivation is to dump. We're considering an airdrop to STEEM hodlers, but we also don't want to airdrop tokens to a bunch of people who are fundamentally not that interested in crypto / will never use the platform. Lastly, we're also considering starting the supply at literally 0, like Steem, and "pre-mining" some for the founders to establish initial influence in the network, and then opening it up to the public and letting the supply grow entirely through rewards.
Thank you again for all of your feedback. We're big fans of DTube, so it's an honor talking to you about this!
We just published an updated version of the white paper that describes the token distribution I mentioned in my previous comment. Rather than doing an ICO, we think we will "pre-mine" a small pool of coins for the Cheddur team to establish initial influence on the platform. From there, rewards will be distributed out to the public via the Rewards Pool (for high-quality content, with the help of our voting) and the Bounty Program (for completing valuable actions).
We anticipate that after the first year of operation, influence on the platform by the Cheddur community will supercede the influence on the platform by the Cheddur team. This sounds like the right amount of time for us to vote and funnel CHDR tokens into the hands of a large enough set of valuable community members.
To accommodate this new design, we also reduced the starting supply of tokens from 100M --> 5M, and ramped up the inflation schedule from 4% starting with 0.5% annual decline, to 20% starting with 2% annual decline.
Would love to get your thoughts on this!