With 10% threshold a part of this curve just does not exist.
I think the point of the graph is that your statement is factually incorrect. And itās super confusing (I was confused for a month) which is why Iām trying to target misconceptions, but not weigh in on opinions. The presence or lack of the 10% barrier does not affect the [RPL/vote power] curve.
Late to the party here. But my decision is easy. Consider someone who runs several minipools and has a decent chunk of RPL staked as a result. Now consider recent price action and its impact on previously healthy collateral. In this instance that individual would still have considerable RPL at stake and a keen interest in governance, but being below 10% at the time of any vote would prevent their voice from being heard.
Being below 10% either by a little or a lot doesnāt detract from the fact they have committed to the protocol and have something (potentially a lot) at stake.
7 Likes
I agree with @NeverAnIsland , even though the voting power per RPL from quadratic voting itself doesnāt change, the effective voting power of an individual voter does change as a percentage of the total voting power eligible to vote.
My main sticking point keeping me voting for >10% is this outsized voting power granted (percentage wise compared to all voting power eligible) to people who keep their node collateralized as low as possible. Itās a perverse incentive that could make people explicitly never top up RPL on their node to maximize their governance power, instead turning up new nodes w/ minimum collateral and letting them get as minimally collateralized as they can (by RPL price decreasing relative to ETH).
1 Like
This isnāt actually true if <10 are allowed to voteā¦
If Iām at 10% and have 25 RPL I have 2.5 power. Say the price changes, and now Iām at 5% and have 25 RPL and 2.5 power. If nobody else stakes or unstakes RPL, we can safely assume the total vote power has increased, as at least one node operator was over-collateralized, and now has a higher power.
Therefor, my percentage of voting power can only decrease, strictly speaking, as I become under-collateralized.
3 Likes
Would you also call rewarding people for exiting some minipools by giving them voting power back a perverse incentive?
1 Like
So I finally made it and I hereby acknowledge that I was delusional all this time about āoutsized voting powerā. It was an interesting mental bug and Iām happy to debug myself, thanks to all people involved.
Iām in favor of āallā strategy now, because its more secure. 10-150% strategy diminishes governance security in times like this because it lowers a quorum significantly.
While āoriginalā intent was probably to support RPL price, I suspect that the same mental bug could be a part of reasoning as well. So in the end we have two main arguments: a greater protocol security on one hand and better RPL buying incentives on the other hand, but as Joe argued recently it is quite questionable. I think security impact is clearer, much higher and is more important.
8 Likes
āBelow 10% have outsized relative voting powerā
Lower collateral % does not result in more relative voting power, lower staked amount does.
The smallest eligible stake amounts have outsized relative voting power regardless of whether the cutoff is at 0% or 10%, this is a feature of quadratic voting and a threshold does not reduce the effect.
āNodes above 10% are more alignedā
There is no obvious link between collateral % and alignment.
Someone with one freshly launched minipool is above 10% collateral, while someone with dozens of LEB8s and thousands of RPL staked for months may not be.
āWhat about Sybil attacksā
Sybiling nodes for voting power results in you having 10% collateral, above the threshold in either case. If the price drops subsequently so that these nodes are below 10%, they have reduced relative voting power as nodes above 150% collateral gain effective stake.
On the other hand, >10% does create a plausible attack scenario whereby nodes are forced below the 10% threshold, removing them from the voting pool, this is not possible with eligibility for all nodes.
4 Likes
This is to illustrate how current 0-10% nodes donāt have outsized voting power.
They are those below blue surface and as you may see, their stake (and hence voting power) is in the same range as the stake of those who are above surface (in 10-150% group).
With RPL ratio going up the blue surface goes down, and vice versa.
1 Like
This is a confusing topic. What makes the voting power outsized is that you shouldnāt be able to have 2.5 voting power for 25 RPL (0.1 VP/RPL) when the best a new entrant could get is 3.5 voting power for 50 RPL (0.07 VP/RPL), as that is the minimum amount they could stake. Allowing <10% to vote allows early entrants with minimum RPL collateral to gain an unfairly high Voting Power per RPL staked.
Allowing this voting only matters when the RPL/ETH ratio declines, letās take it to the extreme. Say the RPL/ETH ratio is now 1% of what it was, now that person with 25 RPL has 2.5 voting power and a new entrant with a minimum RPL stake of 2,500 RPL has 25 voting power (0.01 VP/RPL). So new entrants are voting with 1/10th of the voting power per RPL.
Edit: As an addendum, I would support allowing <10% to vote if the Voting Power per RPL staked was capped at the VP/RPL of the 10% threshold.
This is the tradeoff between RPL collateralization value and total rETH supplied. The same argument can be made against the >10% staking rewards.
First: thanks for posting - glad to see some opining from that end.
Can you help me understand why āper RPLā is the important metric for fairness? In your extreme example, two people join with 2.4 ETH invested into RPL. One of them has 2.5 voting power, and the other has 25 voting power. How is the latter unfairly low?
If folks could choose to enter at 1%, my opinion would be quite different, but as it is the only way to reach that level is to get rekt.
1 Like
Iāve voted in favor of the 10% threshold and would like to share some of my thoughts. Iāll preface by saying that I know some of you have spent way more time than I thinking through this decision than I. Iām sharing my thoughts to explain my vote, not to debate. In no particular order or structure:
-
I think that this vote will ultimately be fairly inconsequential. RPL/ETH is likely at or near itās floor price and weāre unlikely to see new nodes drop below 10% in the future. Rocket Pool validators will continue securing the beacon chain either way.
-
The protocolās current design is built around the assumption that the 10% RPL bond is important for the security of the protocol. Not paying nodes below 10% any RPL rewards is a design decision intended to incentivize those nodes to stake more RPL and return to a healthy status. It is logically consistent to not allow those unhealthy nodes to vote in governance either. There are reasonable criticisms of the necessity of that 10% bond as additional collateral for security but revising the protocolās bond requirements are out of scope of this vote.
-
The other purpose of the 10% minimum bond is to create an economic demand lever for RPL so that the pDAO has more spending power to do things that grow the protocol. Again, there are reasonable criticisms of the current tokenomics of Rocket Pool but those are out of scope here. I welcome discussions around ways to improve the economic design of Rocket Pool/RPL, but working off of the current tokenomics, I think it is reasonable to use governance rights as a way to defend the 10% bond requirement for node operators.
RPL is how we pay for all protocol improvements and growth. Itās not greedy or value-extracting to do things that boost RPL value, it is pragmatic for the long term health and growth of the protocol. Current RPL tokenomics are built entirely around the bonding requirements for node operators. We should either defend the tokenomics as designed, or work to change them in a thoughtful way. Allowing undercollateralized nodes to vote undermines the current tokenomics imo. -
It is the responsibility of the investor/staker to know and understand the market risks before buying and staking RPL. We all know that crypto can be extremely risky and volatile. It is not the responsibility of the pDAO to ābail outā node operators who bought too high and give them governance rights now that they are undercollateralized. Iām not trying to be callous here - I assure you that Iām feeling the RPL price and liquidity pains as much as anyone else out there. I think a lot of the subtext of the discussion here maps back to wanting to appease node operators who got rekt in the RPL market. I imagine this discussion might look very different in a bull market.
-
I have mixed feelings on quadratic voting. While itās not the main reason for my vote, it is certainly something that this discussion has prompted me to think about. I think quadratic voting is very helpful for generating good conclusions from a diverse set of ideas (like grants for community members), but it does introduce a fairly easy attack vector for bad actors. Personally, it does feel a little arbitrary that my voting power is a lot less than if I had split into many nodes instead of just one.
Anyways those are my thoughts 
Iāll be reasonably happy no matter the outcome. I think this is the hardest bear market Iāve faced in all my time in crypto, and I think some of that frustration is getting channeled into the discourse here. Take a breath, keep building. Weāll all make it through 
8 Likes
It strikes me as important for the short term, even if you believe RPL is up-only from here. We have a lot of things to decide, and I donāt think those decisions should be made by an artificially small sample of the node operators. We should strive to pick an option that will weather any market conditions so that we have the most robust governance system possible.
I think the bribery math above shows that the 10% minimum is potential worse for the security of the protocol. Itās entirely feasible that a 10% minimum improves security when itās for rewards, but hurts it for vote power. Is there a reason to think theyāre not distinct here?
I disagree- allowing undercollateralized nodes might make some node operators less likely to top-up, but it may also make some node operators more likely to buy RPL in the first place.
The tokenomics may even be improved. A bond of 10% of borrowed eth, in RPL, will still be required to launch minipools. I see allowing undercollateralized nodes to continue to have a voice in governance as stengthening the value prop of RPL. A token that allows you to participate in governance in a wider range of prices surely has strictly more utility than a similar token with a narrower range.
The framing of this point makes it sound as if itās a given that <10% were never allowed to vote previously. RPIP-4 did not explicitly say as much. I object to the framing of allowing all nodes to vote as being a ābail outā or āchangeā vs the idea that >10 is the āstatus quoā- we failed to establish a status quo.
As such, I donāt see how the pDAO is acting to bail anyone out here. This comes down to whether or not the ideal system of governance for RP includes a minimum collateral for voting or not.
Iām with you on this one, and previously advocated for a switch to linear. That said, with sublinear bonding curves on the horizon, I believe weāll get enough sybil resistance to justify keeping the quadratic curve around.
2 Likes
Maybe fairness is too squishy in this situation. I would point to the recent proposal you put forward for moving RPL staking rewards to a vaguely quadratic system that gives maximum rewards per RPL in the 10-15% range and 0 rewards per RPL below 10%.
Why isnāt Rocket Pool as a protocol consistently incentivizing an RPL stake around 10%? The quadratic voting system is consistent when it requires >10% to vote, however when anyone can vote itās incentivizing the minimum RPL stake attainable.
Voting Power per RPL is important because it doesnāt matter when the person bought their RPL and for how much ETH. What matters is how much the RPL staked is worth in ETH terms currently (especially for slashing purposes).
You think of staking rewards in terms of per RPL so why wouldnāt you do the same with Voting Power?
That proposal scales rewards somewhat with RPL and more so with borrowed ETH (aka rETH supply). The bit where adding another minipool increases rewards is quite different from what weāre talking about in voting. At a very high level - I donāt think the two (voting and rewards) must be similar.
when anyone can vote itās incentivizing the minimum RPL stake attainable
My belief is that the financial incentive is, on average, much stronger. That said, I do think folks can be annoyed to not have a voice. I personally see no benefit in denying that voice, which is why I lean slightly towards āallā (as ever though, my main position is to make sure we have a clear outcome).
What matters is how much the RPL staked is worth in ETH terms currently (especially for slashing purposes).
I agree that for secondary collateral purposes the value of the RPL matters. However, this is most beneficial for would-be-thieves, who we donāt expect to top off regardless of vote power (in other words - the entry value matters). Thereās also the matter of RPL slashing for MEV theft not being a thing right now, but thatās a separate issue.
You think of staking rewards in terms of per RPL
I do currently, but the proposal you mentioned makes it primarily about ETH worth of rETH minted (while yes, retaining an RPL amount term).
It is quite presumptuous to think that the >10% option is boosting RPL value. I would argue it is greedy and value-destructing because it diminishes RPL value.
Could someone assure me that allowing undercollateralized nodes to vote wonāt eventually result in a strong push to pay RPL rewards to these undercollateralized nodes? If not, why not? Sounds like a bit of a slippery slope to me.
2 Likes
Anyone that claims they can know that is experiencing hubris 
The best I can give you is some of my personal feelings:
- Most current value models care most about the relationship between requirement to enter and rETH supply, and secondarily about the incentive to top off (ie, in a mature state topping off beats out not topping off). I donāt think it makes sense to damage that (unless thereās a significant change to tokenomics which changes value capture, ofc).
- There have already been suggestions to give a few types of grace to folks under the limit ā a small ramp down, some time grace, etc. I like the cliff myself, but I donāt think models change much if the slopes are strong (ie again, in a mature state topping off beats out not topping off). These proposals predate discussion on voting body change, so Iām not sure I see much of a connection.
- I donāt think this voting block is overpowering (at all). The evidence? They donāt even have a massive lead in the āallā variant for this vote, which reads roughly as āgive yourself powerā.
- I see no reason these two things need to be the same (in fact, my rpl staking rework proposal changes rewards a lot without touching voting). In other words, theyāre not the same mountain to me, so it doesnāt matter if one of them has a slippery slope.
- I personally would vote against such a thing (that stance was active in rpl staking rework discussion, itās not just pure theory), despite slightly preferring āallā here (though again Iāll emphasize my priority is agreement).
1 Like
I think the discussion has fallen into two camps so far, those who think thereās a strong relationship between the proposed minimum collateral for voting eligibility and the minimum collateral for rewards, and those who donāt.
I, personally, see the idea that theyāre the same as a product of naive design of the voting system. That is, just because we have a 10% minimum for rewards doesnāt mean we need or even should have a 10% minimum for votes. We need to analyze the benefits of each independently instead of copy pasting the parameters from one onto the other.
So it would be incredibly hypocritical of me to follow up this argument (that rewards and voting parameters need not overlap) with an argument to make them overlap at 0%.
Inflation is a tax on those who are not staking RPL effectively, and that tax is disbursed to those who are. I see this as a positive mechanism by which to attract node operators and grow the protocol. I am happy with the cliff at 10% for rewardsā¦ I donāt see why we would financially incentivize a node operator to provide collateral when the collateral they provided is less than the minimum to secure the protocol, as it is damaging to security.
However, I do not see damage to security in allowing them to vote- on the contrary, I believe giving them a voice broadens the voting pool and provides extra security, eg, against a bribery attack or even a whale sybiling the protocol. That is, if someone with a lot of resources wants to attack the protocol, theyāre going to do it by making a bunch of nodes with one minipool each, staking 10.01% collateral, maximizing their power on the quadratic curve. When this happens, would you prefer the nodes who are honest to be eligible to vote, or possibly ineligible due to being below 10%? Our resistance to such an attack is much stronger if everyone can vote.
Following this logic itās easy to see how a 10% minimum for rewards improves security, but a 10% minimum for voting damages it.
5 Likes
Agreed - donāt like the idea of a node operator losing their voice over price dynamics.
2 Likes