RE: Steem Witnesses - Weighted votes based on ranking
Great to see this more in-depth and systematic breakdown of a governance change proposal!
The section that was of most interest to me, How much stake would be needed to control the blockchain?, was not too clear to me. I could not understand how the ranking system would be different - how would it require owning a less amount of stake (than our current governance mechanism) in order to control the blockchain? It seems to me that if there are two stakeholders (a single entity and the community), then the ranking rules would apply to both. Any changes made would not put the single entity at a disadvantage while at the same time being neutral to the community's stake, right? Or maybe I can't envision how the ranking rules would only put the single entity at the disadvantage and the same disadvantage not being applied to the community (effectively cancelling out the change).
Weighted ranking puts a single account at a disadvantage if the intention is to control the blockchain. This is because their stake only significantly impacts a few witnesses. The community on the other hand would be ranking many witnesses between 1 and 20. This would result in less weight on the top few witnesses but more weight on the lower witnesses.
In the above example, the single account would control about six witnesses compared to 14 by the community even though in the example the single account has equal stake to the community.
But the point is really about a single entity, not a single account. A single entity can divide its stake across multiple accounts. If the single entity divides its stake across multiple account, is it still at a disadvantage (while the rest of the community not)?
Also, I am not completely clear on what the percentages refer to in the above diagram. Some more labeling of the variables you include in your diagrams will help as currently it's not very clear. But I really appreciate you making the diagrams as it helps put the conversation on a solid foundation.
I think the key question is how many witnesses can one entity control. Dividing stake into more accounts could increase the number of witnesses that one entity can control if that entity has sufficient stake to do so. In the current system dividing stake is not necessary as each account has 30 equally weighted votes. Reducing the number of votes or reducing the weight of votes based on a rank would likely offer incentive to an entity to divide stake between accounts to obtain more witnesses. If the number of votes are reduced sufficiently or if the weighting is heavily skewed, it becomes close to impossible for one entity to control the required 17 witnesses as their stake supporting witnesses is greatly reduced by dividing the stake into several accounts. I explain in detail in another post using an example of reduced number of witness votes.
Yep, I get that. And I read your post which was again well detailed.
The downsides that I see from reducing the number of witness votes are ones which you have also mentioned:
On a positive note, Steem is becoming less susceptible to centralization as newly created Steem is getting distributed to more users. It is also difficult to buy a sufficient amount of Steem from exchanges to have a strong influence on the blockchain. The only way to do it is to buy large accounts such as the Steemit accounts. Even these accounts hold a considerably smaller proportion of stake compared to what they held a few years ago.