Proof Of Worth: Revisited And Upgraded

Three months ago, I proposed a possible solution to self-upvoting: Proof of Worth. The idea behind proof of worth was that in order to verify one's self upvote, the sum of the non-author upvotes needed to exceed that of the upvote itself. This solution would prevent lone whales from upvoting garbage and stealing from the rewards pool. This proposal also preserved the self-upvote. However, one could simply argue that one simply needed an upvote from a different account to bypass this idea completely. So, I extend the idea to make it more useful.

Proof of Worth: A natural extension to proof-of-brain

---

Proof-of-Worth vs Proof-of-Brain

Before we dive into the more technical details of the new and improved Proof-of-Worth proposal, I want to look at things from a more philosophical perspective. Proof-of-Brain rewards are granted to individuals that create content that is then judged to be worthy of rewards by at least one person. Proof-of-Worth rewards are the same. But Proof-of-Worth rewards are more restrictive. In a sense, Proof-of-Worth is a subclass of Proof-of-Brain.

The restriction put in place in Proof-of-Worth is on the validity of members of the community. Something that has massive rewards but lacks widespread support would fail this higher standard. The rewards placed on any post should reflect the amount of support the community as a whole has given that post.

A whale vote that makes up the majority of rewards says less about the community's support of the post than the remaining votes that surround it. Thus, this whale vote should be penalized such that the total value of the post is more indicative of the value of that the community places in that object rather than the value assigned by an individual or a small group of individuals. Proof-of-Brain does not take this into consideration and thus the reason we have fighting over farming the rewards pool.

Simple Proof-of-Worth

In the case of simple proof-of-worth we take a look at only the largest vote among a set of votes. This vote may be a self-upvote or it may be the vote of another account. In this case, we look to avoid an undeserved massive single upvote. In our case, "undeserved" means that a vote lacks community support to verify that vote as representative of the community sentiment of that post.

Simple Proof-Of-Worth: In order to receive the full rewards from the largest vote, the sum of all smaller votes must be greater than or equal to the largest vote. In the case where the largest vote is greater than the sum of all smaller votes, the largest vote is penalized such that it composes only 50% of the total rewards allocated to the post.

So, for example let's say a post has five votes on it by a group of wealthy Steemians. There are $310 in rewards on the post. But when sorted the individual vote totals are $300, $4, $3, $2, and $1. Apply our proof-of-worth restriction, the total rewards are now $20 instead of $310. And this appears to be a fair value for this post considering it only received 5 relatively large votes.

What if a post only has two votes? Well, then the value assigned to the post is the smaller vote multiplied by 2. What if a post only has a single vote? Well, then then the value assigned to the post is 0. Thus, if you self upvote and receive no community support, your post is deemed worthless.

This solves the problem of individuals upvoting their own shitty content and it solves the problem of a single whale upvoting mediocre content excessively. This also helps to curb other voting abuses. To abuse this system to same extent, one would need multiple accounts and twice the amount of money in order to compensate for the proof-of-reward restriction. If one did do that, the price of Steem would go up. So, win-win-win.

Extended Proof-of-Worth

One may ask: why simply validate the largest vote, when we can validate all of the votes? Extended Proof-Of-Worth extends the idea to reduce the impact of larger votes than normal votes in general. First, we need to pick a starting point in our list of sorted votes and then apply simple proof-of-worth to them individually comparing them to all of the smaller votes. We'll use the median as a starting point as it is more resistant to skewed distributions than the mean is. And we are trying to temper the effects of these skewed distributions.

Extended Proof-Of-Worth: Sort the list of votes in ascending order. Find the median value of the votes and pick the smallest vote equal or greater than the median value. For this vote to apply completely, the sum of all smaller votes must be greater than or equal to this vote. If this is not the case, reduce the value of the vote so that is it equivalent to the sum of all smaller votes. This reduction is applied after the entire process is done. The old value will be used in the remaining calculations. Now proceed to the next vote in the list and repeat the process until the largest vote is evaluated. The new rewards total is the sum of all votes after penalties have been applied.

With the example we used earlier, the final total is $20 and the same as simple Proof-of-Worth. The median vote in this case is $3. But $3 is equal to the total of the smaller votes so it is not penalized. We then move to $4. It is smaller than the proceeding votes so it maintains it's full value. Then we reach our final vote which gets penalized the same as we are now in a simple Proof-of-Worth case. Extended Proof-of-Worth simply extends the simple Proof-of-Worth process to smaller votes.

Let's use another example: Let's say three whales collude to upvote a post. We have seven votes this time: $1, $1, $2, $5, $80, $100, $120. Normally this post would be $310. If we apply extended Proof-of-Worth, three of the top four votes are penalized. $5 becomes $4. $80 becomes $10. $100 becomes $90. In this case, the top vote remains the same. In simple Proof-of-Worth, this post would still be $310. In extended Proof-of-Worth, this post is now $231. This is still high, but it helps to somewhat alleviate the issue. Even though whales are colluding here, 3 whales is more representative of a community than say 1.

Note that in the case of posts having one or two votes that extended Proof-of-Worth is the same as simple Proof-of-Worth. This also helps to alleviate some of the potential issues of a user abusing multiple large accounts.

Potential Concerns

One may asks what happens to the curation rewards if a vote is penalized. In the case of the last example, it would appear that the owner of the $80 is a really big loser as his contribution to rewards is only $10. But their proportional allocation of the curation rewards would still be the same. Their share of the curation rewards is preserved while the amount of curation rewards would simply be a little smaller. Thus, the penalty only serves as a very soft cap on the rewards allocated to the author (and curators). Which makes sense. We are trying to give out rewards that the community can agree upon.

Comments tend to receive significantly less votes than posts. Thus, this solution may seem to disproportionally affect users who comment a lot. There are two things to mention here. The first thing that we could argue is that only the truly exceptional comments deserve rewards. The second thing that we could argue for is by placing an arbitrary lower limit on votes that proof-of-worth would apply to. Something like a 1000 SP vote. Which approach is taken is up to the community to decide. Penalties on larger votes would also stop at the cap.

The potential for reduced or lost voting power is another concern. By using the cap suggested in the previous paragraph, no one will ever suffer a complete lost. But the potential for bad ROIs should encourage larger users to use their votes more carefully and spread the wealth more horizontally than vertically, since they may lose out on curation rewards on "excessive" votes.

One last concern is that of flagging. If we use the same system for flags and upvotes, flags by whales become super-powerful. Thus, in order to implement the system, we need to separate upvotes and downvotes from each other and apply proof-of-worth on both lists. In the case of negative upvotes, the list is put in descending order rather than ascending order. Or you can use the absolute value and keep the ordering the same. Whenever makes more sense to you. With this system, we now need multiple users to flag to punish a post. This prevents a single user from flagging other users and limits the damage of lone whale flags.

Final Thoughts

I really like this idea. The first time I came up with the idea, I found it rather weak and not through enough against abuse. This new solution is more through and preserves the use of the self upvote. Apparently people like voting for themselves. In this system, they still can and use that upvote to promote their work. If nobody likes it though, they receive nothing, which seems fair. In fact, by adopting the system, we can finally begin to lose the stigma against self-voting as most abuse is curbed and the most egregious abuse should die out completely.

The final thing I have to say comes in the ask for a small favor. I embrace constructive feedback, but often flaws are used as reasons to strike down an idea. Thus, in addition to any criticism or praise, I'm just curious whether or not you would prefer a Proof-of-Worth type system over the current system. That being said, you have all the freedom to do whatever the hell you want to do. At a minimum, I appreciate the time you've taken to read over my proposal. Thank you.

Sources

Image