Here I offer a proposal to improve curation, increase variety on the trending stories, and promote easier discovery and rewarding of new authors.

img source

Steemit curation mechanisms provide very hard incentives in form of curation rewards to vote for things that are bound to become popular with certainty. That fact drives user behavior through follow lists of popular authors, voting often without reading the content, mutual group voting and automated curation through bots and other tools.

Some projects like @curie look to shift the scale a bit towards the discovery of new authors, but a lot of effort is spent against stacked odds when established authors get a huge amount of rewards in a few minutes no matter what they post.

I understand this proposal may be met with resistance from these established authors, but I firmly believe that a more varied platform with a lower barrier of entry is good in the long run for everyone.

Repeat votes from the same person to the same author should weight less

My proposal is simple: repeatedly voting for the same authors should have the effect of each subsequent vote weighting less.

The numbers should be analyzed for balancing, but the best way I can think to accomplish this is the following:

When voting, count the number of votes given from this voter to this author in the last seven days. Each vote halves the effective voting power used in the vote.

So for example, if you voted once for this person in the last week, the next vote will be cast with 50% voting power. The third one with 25%, the fourth with 12.5%, and so on.

Multiple Benefits

Not only this helps promote the curation of new authors increase the variety, it has a few, very beneficial side effects:

• Bots are negatively affected: automatically voting for "sure things" is not a good strategy anymore.
• Voting rings are negatively affected: anyone controlling a large number of accounts attempting to grow them by automatically voting on each other's content will have a much more difficult time.
• Serial flagging is negatively affected: there have been a few cases of revenge and automated flagging recently. Flagging multiple posts from the same author in a row will have less effect.
• There's no negative effect for the voter or author: the vote simply uses less power, which stays available to be spent in the next vote. Voters don't really need to think and keep track when they last voted for someone unless they are trying to game the system.

Having looked at the voting code, this seems to me very viable to implement, could have an immediate effect on cutting down bots, increasing variety of the trending content, create new niches and lowering the barrier of entry to the platform.

What do you think?