Lots of discussion points appearing on here and on Discord! That's great to see. Just hopped out of a call with @hudsonsims covering a lot of points presented by @realbtt, @rarest and @pepe.
Firstly, for amendments to my initial post, I would propose the following:
Linear VP Regeneration
Logarithmic implementation of VP has been provably sybilled in the past. Pepe sums up the methodology aptly here:
pepe VP generation has created individual decision power over what gets minted for a few single individuals that have played by the flawed rules of 1 user = 1 staked address (current system). With less than $100 staked in total across a bunch of addresses any of those actors can pretty much sway any weekly vote to their desired outcome. This is actively happening today. So not only does this logarithmic design disincentivize large holders past a tiny amount, but it works against the small and less informed holders, who mostly stick to one address.
Rarest's thoughts on Discord echo this:
VPs should be directly proportional to the amount of BOTTO staked. The logarithm in the VP formula to limit the influence of “whales” is a huge mistake as it disincentivizes buying and holding BOTTO tokens, and it over-complicates something that when it is simple works just fine.
This encourages a return to the initial implementation of Botto which has justifiable effects on the protocol and token re. economic incentives. Specifically, incentives to acquire more $BOTTO over time increases so as to increase an individual's own sway on the artist. Coupled with the BB&B mechanism initially introduced and reintroduced in my (tentative) proposal above, $BOTTO will look a more attractive token to stake and hold.
Some comments based on @pepe's post:
pepe Voting must happen 1 token = 1 vote for the Artist to survive. It’s the simplest and most elegant solution. A few of us in the community have advocated for it since the beginning. That’s also how Botto BIP governance works already - and no issues at all there.
If you mean this by as being represented by VP, i.e. just defaulting back to linear VP regeneration, I agree.
pepe Because of its flawed fundamentals, after 12 months current VP distribution is already too far wrong to be fixed.
This is not true. VP is offchain, and can be reset (we don't want to do this), or can roll over (as initially intended) into the new period. VP will be generated incredibly quickly to the point where VP distribution will be normally distributed relative to staked amounts. I'd say the difference would be irrelevant after a week of VP regeneration. You'd also see people with 5/6 figure $BOTTO generating ridiculous amounts of VP, completely nullifying the idea that this is unfixable.
pepe We should abandon VP generation entirely.
If you have a clear proposal on how to move forward with this, I'd be keen to understand what you think this looks like! For example, If we implemented a pure 1 token = 1 vote mechanism, how can we vote across multiple weeks, rounds, periods? I am uncertain how this looks like but would like us to explore winning implementations, whatever they may be. Linear VP regeneration is IMO a first clear step without radically overhauling systems we currently have in place and is correct progression for the second period.
There are other, trickle down effects on this proposal should we move to linear VP regeneration:
VP Slider will have a range of 1 to custom defined value
This change straightforwardly encourages uncapped voting behavior and is required to allow users to spend their votes proportionately to their VP regeneration.
Removal of Social Verifications
Sybilling becomes a lesser issue with linear VP regeneration as there's no incentive to have multiple accounts. We can omit this from the proposal.
A trickier discussion we should further explore is how to reward behavior actively (Active Rewards contract).
Active Rewards (25% Revenue Redistribution) - IMO the biggest thing we need to discuss
There are two schools of thought here: either reward voters linearly proportionate to their VP expenditure in any given round, or alternatively reward voters linearly proportionate to their VP expenditure on a winner in any given round.
The obvious default to this would be to reward users based on their VP expenditure in any given round. The core issue with this is that voters will mindlessly vote, without consider choice, on the fragments in front of them. There is no element of collusion, and less element of requiring any thought to be put into voting behavior. This is a message that might be misconstrued. IMO Botto is about art, it is not about (blind) economic incentive. Naturally, there should be economic reward for positive behavior (training Botto, helping curate a winning piece to auction). As a protocol we need to ensure that voters are mindful of their vote allocation and that vote allocation remains meaningful. At the same time, I see benefits to linear distribution proportionate to VP expenditure. It's inclusive, easy to digest, and provided we individually understand the best artwork must win for a payout, it might be the right implementation.
Nevertheless, I'd still suggest that we reward those who voted on a winning fragment proportionately to their VP expenditure. This requires a more active version of considered thought in outcome, social collusion, and still offers potential inclusiveness in receiving a share of reward should that be the only thing the voter cares about. This also encourages earlier curation in the voting week for voters of all sizes to mindfully curate artworks capable of selling at auction, as early week votes will have direct impact on frequency of appearance in the voting pool later in the week, and thus subsequently what appears on the leaderboard.
The first argument against this from rarest:
Rarest: Also, rewarding only those who voted for the winning fragment doesn’t make sense, as big holders could collude and vote for the same fragment. It makes sense to just reward all those who voted, proportionally to the VPs spent, and regardless to the fragment they voted for.
Social collusion is in my opinion important, where big holders can do that anyway irrespective of how users are rewarded. There's no guarantee that larger players would agree on what to curate anyway. By rewarding voters who spent VP on winners provides the added incentive of curating earlier (influence what others see later) and curating with purpose (not just blindly dumping VP to earn a piece of the pie).
@pepe is not seem fully convinced by the idea of redistributing to those who spent VP on the winner:
pepe Additionally I’d say that the division proposed (voters on the winning piece get 100% of rewards allocated to this, voters on the rest of the pieces get 0%) is a bit drastic, given a lot of the initial curation is done by the people who vote pre-Leaderboard, and in a 2-at-a-time show you might not even see the future winner. Also you would be incentivized to vote what most others are voting, not what you think will sell best. Perhaps distributing to all voters equally regardless of the pick, with 25% to pre-Leaderboard and 75% to the last 24h could work as intended.
Initial curation done by people who vote pre-leaderboard has a direct influence on voting activity later on. Most VP expenditure comes from end-of-round boosts. Users are able to influence what appears on the leaderboard simply by curating in the early stages in the round. I am unsure whether larger voters would vote solely on what others are voting. With the leaderboard being a blind top 15, how are users to know what's being voted on?
That said, I suppose the effect of rewarding winner-only could disincentivize curation pre-leaderboard - I don't think any of us are certain what the effect would be. Finally, the DAO would still have 25% of rewards to distribute retroactively based on what we deem positive behavior (this could also be used as justification for simply rewarding users based on VP expenditure too).
I think we have two options here with significantly different tradeoffs, I'd be keen to hear from everyone which one we should opt for and why.
Payouts in ETH
I think there's enough reasoning in place to have payouts denominated in ETH in both active and retroactive rewards. This removes any form of buyback for active and retroactive rewards.
Qualifying for Active/Retroactive Rewards - 2000 $BOTTO
This imo should remain, see my comment above