Age and Activity Improvement Proposal

co-written with @vern

This is a draft for a proposal that outlines adjustments to the age and activity coefficients used to calculate influence, and to the algorithm for normalizing activity. Currently, a user’s activity can be negative if the total rewards they’ve received is less than the total value of their votes. As a result, some users have negative influence, and therefore are unable to receive rewards. The protocol was intentionally designed to penalize users for doing a large number of low value votes, but the existing influence algorithm is overly punitive.

Therefore, the influence algorithm should be adjusted in the following ways:

  1. The relative weight of age should increase from 0.33 to 0.4, and the weight of activity should decrease from 0.67 to 0.6.
  2. Activity, when factored into influence, should have a minimum value of zero. This guarantees that no accounts will ever have a negative influence while ensuring accounts with low value votes are still penalized.

Will post to Snapshot tomorrow after feedback.

18 Likes

The second point here should address negative influence @cryptoleon @Dr.xavi @Inter_netMoney @GabyCarfe17

4 Likes

Hi nir! Thank you for sharing your proposal. I was agree when PBJ told me that increase the weight of the age of the token was on your plans, and I still believe that it’s a great idea.
About the second point I’m totaly agree, it defenly will fix the problem with the negative raw influence. Looking forward to this changings.
Regards!

6 Likes

I am in agreement with this as long as allows a dapp user to change the condition of their influence and not remain or go to far negatively where the account isnt even worth using anymore

4 Likes

Can you explain “low value votes”?

2 Likes

Yes, the @harsh question is very important: can you explain: low value votes?

Because I wonder if they are going to continue penalizing and lowering the influence to those users who vote content that nobody has voted and maybe in 24 hours, 1 week or 1 month nobody votes?
Because it is enough that if I vote something and no one votes I do not receive rewards.

The system will be designed to reward those who vote for popular content by the order of arrival?
Because I spent time discovering and curating NFTs with YUP when no one else had done it, I did it with the NFTs I liked and since no one else voted for those NFTs I found that the value of my vote was higher than the rewards I received and I was penalized with negative influence while there are people farming day in and day out and they are still rewarded.

In short YUP protocol will reward the content curators or are we all going to farm YUP to get rewards and not be penalized for discovering new content that we like?

I think they should make it very clear how the voting and reward process will be once they make this change.

8 Likes

exactly what I need to know or account is in this condition I dedicated myself to voting on different things that I liked and now my influence is negative

6 Likes

low value in this case refers to the return on the vote. A vote that isn’t followed by any votes by others can be considered a low value vote.

1 Like

votes that aren’t supported by others will still not be rewarded, though the time horizon in when that happens is an interesting point. We’re open to hearing other mechanisms for rewarding individuals, but rewarding every vote regardless of its popularity would be easier to game than the current model.

2 Likes

this will be less extreme after this change.

When this change Is gonna apply? To start voting again with no problem for the influence and allow ppl to get rewarded @nir

im ok with this. As i understand the first vote on popular content gets the reward. This is what I understood the protocol to do all along. Hence why I have a high dapp score and a fair ample amount of yup tokens… oddly my raw influence is negative. However I am ok with this feature.

1 Like

OK, I hope that this is implemented in the protocol and the negative influence can be solved, a sadly that I do not have Yup in ETH to participate in the vote.

Don’t you think this will somehow create a habit of voting only popular content? Not the content that actually is good.

We have discussed many many times that people should vote the content that they think will be popular in future. I vote content that I genuinely like, not the most voted tweets. People just want rewards and they will not vote same as me and that resulted in no rewards for me. I saw on chain data of yup rewards. People who are getting loads of YUPs are the ones who are voting popular posts of Elon or Binance or Yup. That’s all.

“Low value votes” contradicts the vision of Yup curation.

Those votes should be considered as low value votes which are most common, most crowded and follow-the-sheep kind of votes. That will force users to vote on more other things.

I hope I made my point clear.

3 Likes

Please give a strong alternative. How are you measuring good?

For example, majority of yup users vote on Elon’s tweets. Does that make it good content? No, because Elon’s activities are majorly voted on Yup for the YUP rewards.

If I watched a video explaining how can I find next rare NFT. I rated it with heart and bulb. Which in my opinion, a good content. But other people don’t think it will be heavily voted like Elon’s tweets and they decide not to vote same as me.

These 2 are very very real examples showing that single votes does not mean they are less valuable.

I suggest decreasing rewards for such content. Not all at once. But in this case Elon’s tweers, CZ’s tweets, Binance’s tweets are major hub of voting.
That will ultimately force users to diversify their votes. But this is not very good alternative.

Hey @nir I just post this idea:
Although I do not know the technical aspects, I still think that if 10 people vote on a tweet of Elon, 8 of those people rated the same content on the day the protocol should identify that coincidence and penalize them with low rewards, if they have done the same day after day those users should be receiving less rewards to nothing, instead those two people who voted the tweet of Elon and has no high coincidence of votes with the rest of the user should not be penalized.

This way if you remove the money factor in those popular accounts I think people would vote for the ones they like and avoid a pattern of making money that are the accounts of Elon, Saylor, Pomp, CZ binance…

But it should not be generalized, only penalize those who are dedicated to vote the same as others, because if you have 20 votes to 40 votes and 40 users vote the same day after day, or are the same person or are organized, because each person has different tastes to have high coincidence in their votes.