Witnessing

In many cases, moderators would have to vote on some content outside of Questfall, which often can be edited by the author at any time (e.g. a post on Facebook). And this is a problem.

This applies to all external content, because even if a platform does not currently allow content editing, it can be changed at any time.

A user can submit a completion and then edit it while it is being moderated. As a result, the consensus result should change on the fly. This means that the first moderators who voted on the original content would be punished for doing their job correctly, which is obviously not fair.

This may even happen without malice. For example, there may have been a mistake initially, and a user fixed it later, after moderation had begun.

Penalties, however, are a key part of voting protection and cannot be removed, otherwise any kind of random voting will become a profitable strategy, and community moderation would not be able to play the role of Sybil defense.

To solve this contradiction, Questfall implements the concept of witnessing, which implies that users provide two types of information: a link to the content and a screenshot of that content.

Although we implemented this approach specifically for quest completions, it can be used with any external content.

Moderators would first vote to approve that the screenshot represents the actual content posted on the third-party platform, and then make judgments, such as whether or not the quest was completed, based on the approved, immutable screenshot.

Screenshots will be stored within the Questfall infrastructure to ensure that no one will be able to alter them after they are uploaded.

These two steps, witnessing a screenshot and then judging its content, are actually two separate voting topics, and are voted on by different sets of moderators. But they are still two sides of the same coin, and represent the same issue, meaning that the amount of witnessing and judging in the system will always be equal.

Therefore, if there are many such completions, the system would be able to balance such split topics for each moderator individually over many iterations.

So that out of 100 votings, 50 are witnessing and 50 are judging.

The main advantage of splitting an issue into two voting topics is that the incentives can be split as well - while the reward portion goes to witnessing, the penalty portion remains for judging. And since this approach does not penalize witnessing, users would not be able to punish good moderators by changing content.

In short, witnessing allows moderators to earn, while judging allows them to not lose.

However, witnessing is on average profitable in random voting. Therefore, the average profit from witnessing should be outweighed by the penalties in the judging moderation step and the cost of bypassing. This makes it impossible to earn by randomly voting on these two alternating topics over many iterations.