Hello Steemians, I’m @vandeberg, Senior Blockchain Engineer at Steemit and today I want to do a more technical deep dive in to the proposed downvote pool in the Steem EIP. One of the core tenants of Steem is the belief in the wisdom of the crowd to curate and reward good content. The current economic model limits every account's ability to earn by a resource called "voting mana". Whenever your vote, whether it be an upvote, downvote, or changing your vote, some of this voting mana is used up. When it is all gone, your votes no longer impact rewards.

Of these three actions, only one of them rewards the user, upvoting. Neither changing your vote nor downvoting can reward you. It then makes sense that if you greedily optimize your return on investment, you would only upvote content as downvoting it would be a waste of your precious voting mana.

However, downvoting is an integral piece of the curation process. An ideal solution would incentivize downvoting with rewards, but we have yet to come up with a solution that is fair and not exploitable. In the meantime, we believe allowing users to have some downvotes without consuming their voting mana is a reasonable solution. While it does not incentivize curating through downvotes, it removes the direct cost of downvoting, which should make downvoting a more economically viable option.

What we propose to do is to create a separate downvote pool that can contain its own mana up to some percentage of the mana that the upvote pool can contain. Downvotes will be taken from the downvote pool first, and then the upvote pool once the downvote pool has been consumed. The downvote pool will follow the same rules as the upvote pool, regenerating over five days and filling instantly and proportionally to new Steem Power and delegations.

We think having a downvote pool that is 10-25% the size of the voting mana pool would serve as a good starting point. The good way to think of this is that a certain number of downvotes are free before you are charged for downvoting. Charged only in the sense that you are losing potential rewards you could have gotten from upvoting. That number needs to be high enough to make a difference, but not too high that it becomes exploitable.

The obvious alternative solution is to have two entirely separate pools for upvotes and downvotes. We believe this is a bad idea because it would allow the reward system to devolve into a zero sum game without consequence. Each account could award and remove an equal number of reward shares to content. If everyone did this, then no content would have any reward shares and would then not get any reward.

Furthermore, there are users out there that understand this change is adding resources for them to use and will use whatever resources they have available. A downvote against someone else is a small upvote for everyone else. We expect some users will use all their downvotes to maximize their returns. In an effort to curb that behavior, we are recommending to not create a full separate pool.

This hybrid approach captures the best of both approaches. It does not give too many additional resources to users that will use/abuse all that we give them and frees up normal users that may not be downvoting to do so without financial penalty.

Let me know if you have any questions in the comments section below, or if there is another aspect of the blockchain that I should explore next.