The Double Edged Blade of Converting from Staking Based Rewards to Voting Based Rewards

Introduction

I’d like to start by clarifying, this post is not intended to “bash” anyone - I am simply making an observation. I am aware that the proposal being discussed introduced a solution to spam: being the inability to artificially increase the voting rewards of a given day.

However, this has seemingly come with unintended & unconsidered consequences - being the dependency upon Known Neurons & in turn centralization of voting power.

Please also note that the following topic only references Governance topics, as 99.99% of voting power was assigned to DFINITY by DFINITY prior to this change being made.

Resources

Background

In February of 2022, the concept of public “Known Neuron’s” was introduced, offering the community the opportunity to highlight individual neurons within the NNS UI as reputable followees. This, of course, was beneficial to the ecosystem - it allowed users to pool voting power behind those they believe would accurately represent their best interest within governance.

Shortly after, in May of 2022, the concept of converting from Staking Based Rewards to Voting Based Rewards was brought to light - aiming to diminish the financial incentive associated with spam proposals, while simultaneously decreasing the inflation rate of $ICP.

However, this proposal was seemingly shut down - although I can not find the direct proposal from the NNS to verify its results.

The topic then resurfaced in September of 2022, at which point it passed via the NNS, to then be implemented.

As a result, Neurons who do note vote, do not receive rewards.

This meant Neurons that had staked over the previous year, utilizing a set & forget mindset, were left with 3 options to reclaim their APY, had they even noticed the adjustment that stripped them of their rewards;

  1. Establish yourself as an expert on voting topics & vote on every single proposal
  2. Pretend to be an expert on voting topics, vote on every proposal & hope you don’t cause serious harm
  3. Assign your voting power to someone else who has deemed themself an expert & votes consistently

Or in simpler terms, you either lose your APY for assets staked in the past, locked for years in the future, you spend hours a day vetting & voting in proposals, or you simply assign your voting power to a Known Neuron.

As so frequently stated within the Periodic Followee Confirmation topic, most people within the NNS practice a set&forget mindset, attempting to maximize rewards & minimize input - which seemingly resulted in most neurons that were aware of the switch in reward distribution, assigning their voting power to the few Known Neurons available at the time.

Which leads into the question:

Had the switch from Staking Based Rewards to Voting Based Rewards not occured, do you believe we’d still see such a high concentration & centralization of voting power, or is the current centralization of the network a result of the psychological affect of potentially losing rewards?

3 Likes

Hmmm…Maybe? If people more votes weren’t cast, wouldn’t those who had already followed other NN have more say still?

I think the devil is more in the detail of how we ended up with so few NNs. I’m starting to see that DAOs take the path of least resistance most of the time, so UI makes a huge difference. There COULD have been a UI that listed ‘most reliable voters’ in descending order for people to choose from…but then this would have concentrated votes in those reliable numbers.

I think maybe you’re dancing around a fundamental undercurrent of Liquid Democracy which is that, like labor specialization in the economy, you want specialization and then people awarding their vote to those that specialize. I’m not sure we’ve had the best specialization outside the foundation. This is likely a symptom of the price action over the last couple of years.

There is probably a solution that would have solved spam and not concentrated voting power, but hindsight is 20/20. The best we can do now is fight to make the best decisions we can to move to the world we want.

Alternate take: When DFINITY removed themselves from governance, most everyone else remaining was an idiot(no in a bad way…just that no one has really run a DAO like this before…so you know…we were all ignorant). So didn’t most people reselect dfinity for the Governance topic? But some didn’t, so diversification happened…but maybe that wasn’t a good thing. :joy: :man_shrugging:

1 Like

Thanks for addressing this thread so promptly as the original author @skilesare - I want to start by saying I appreciate the governance contributions you’ve made to date, and this isn’t meant to diminish any.

That’s a fair observation! I think the key difference from the scenario insinuated within the article would be the choice.

For example, in the scenario you provide, in which we don’t switch to voting based rewards - those who chose to follow Named Neurons undoubtedly did it not to secure APY, but rather because they trusted them with their voting power.

I believe it is important entirely different when people are borderline coerced into followees.

I think we generally agree here - aside from select topics like Governance or Replicas, we have seen no dedicated named neurons, specifically to code based topics, aside from DFINITY.

I’m starting to also agree that compensation comes into play in this regard, and have been exploring different methods of fairly rewarding Named Neurons (controversial, i know!:astonished:).

Most prominently, I like the concept of drawing similarities to traditional staking/mining pools - in which a minor percentage is deducted from the “earnings” of the pool to compensate the “operator”.

To go a bit deeper; I believe 1% of rewards of voting power cast by named neurons can fairly compensate the operator, while offering negligible deduction from the individual neuron rewards themselves.

However, there’s a reason I haven’t published this proposal yet - this is but a concept that is still being worked through.

I completely agree here, hence why I referenced that my goal is not to bash anyone. I’m hoping this leads to a productive conversation that produces action.

Ideally, to be completely transparent - I hope this sparks discussion on a followee reset again, when considering how we arrived in this position to begin with.

Alternatively, an abstain option can be added.

2 Likes

The spam prevention proposal (proposal 80970) didn’t change the total number of neurons that received rewards. All neurons that voted on Governance proposals before the change still received rewards after that change. However, it changed how the rewards were distributed to voting neurons. Previously rewards were distributed proportional to total votes cast and the change was for rewards to be distributed proportional to total votes that exist.

Voting has always been required to receive rewards, so the spam prevention proposal didn’t change that detail either. I think the relevant change to your point is proposal 34485, which included a change to default following for proposals on the Governance topic. When that change was implemented, the Governance topic was no longer included in the All Topics catch all category. That’s when people had to manually configure their neuron to follow someone they trusted or vote for themselves. At the time, only about 45% of the voting power in the NNS made this change, which is fundamentally what incentivized the spam.

However, the change did result in a greater good because it decentralized voting power on Governance topics by incentivizing people to follow neurons other than DFINITY.

That’s a fair correction! When I originally read this I seem to have missed the specification that the rewards are distributed based on the neuron’s share of total voting weight of the day - completely on me. As a result, there definitely seems to be a misconception in the outcome of the proposal.

Thanks for making that observation before this thread continued too far!

1 Like