The Reputation System: Rewarding Interaction
To maximize quality and fairness, Mindplex will implement an AI-assisted reputation system.
Interaction will be the basis for all rewards.
There are two types of interaction: 1) Interacting with Content, and 2) Interacting in the voting process (Content Request, Governance):
-
Each Interaction will have a Positive or Negative value.
-
Interaction with Content will be the following activities:
1 Commenting 2 Voting on a piece of content via Like or Dislike 3 Sharing 4 Reacting with emojis 5 Consuming Content (measured by time spent on the content page)
-
Interactions in the Voting Process are divided into two parts: Governance and Content Request.
- In the governance voting process, users have three options when they cast their votes: Yes, No, Abstain
- In the Content request process, users can only vote Yes or No. However, users can back their votes with MPX for emphasis. If the request passes, this MPX then goes to the content creator who fulfilled the request.
Interacting with content has varying effects on the user’s reputation, depending on the amount of effort required. Commenting is the most valuable interaction, while content consumption is less valuable. Voting on governance issues is more important than voting on content requests.
Mindplex gives every user the chance to get rewards through these interactions, and every interaction has a result. The wrong kind of interaction will result in a negative reward, which will result in a reduction of the user’s reputation points (MPXR).
The reputation AI will also analyze the sentiment of interactions, and the MPX reward for users will be determined by the nature of their interactions; too many optimistic or pessimistic interactions will result in relatively lower rewards, whereas a rational balance results in a satisfactory reward. When a user’s account becomes inactive, it gradually begins to lose reputation. This prevents “account farming” and the undue influence of inactive users in governance voting procedures.
A media creator’s MPXR balance (their reputation) increases when their content (or a comment they have made) gets a positive rating. The degree of increase is proportional to the reputation of the rater.
Rewarding predictive raters: Being a good “predictive rater” also boosts one’s reputation. A person who consistently gives early positive or negative ratings to content that later receives a large number of positive or negative ratings in the same direction is valuable to the platform because they provide early indications of whether a piece of content is bad or good. Every user’s history of predicting content is kept track of, and those who do this well will see their MPXR balance go up. The extra MPXR tokens are generated in the backend, off-chain, using ‘Prediction Points’. These prediction points are then multiplied by the user’s reputation points for that specific interaction. Only Comment, Like/Dislike, and Reactions receive prediction points.