Variable minipool size w/RPL stake based on protocol ETH

There has been significant discussion in #trading about how best to enable smaller minipools without damaging the value of RPL.

The critical perspective change that folks had was that RPL should be viewed as (A) insuring the protocol’s ETH and (B) giving access to commission on the protocol’s ETH.
Critically, this means RPL should scale with the protocol’s ETH, and not with the NO’s ETH contribution.

What’s been proposed is allowing a range of sizes, with a minimum RPL stake of 10% of the protocol ETH. For a minipool with 16 protocol ETH (and thus 16 NO ETH), the minimum would then be 1.6 ETH’s worth of RPL. For a minipool with 26 protocol ETH (and thus 6 NO ETH), the minimum would then be 2.6 ETH’s worth of RPL.

The idea is that this is a win all around:

  • The smallest possible minipool is much smaller for those that want that
  • Folks that run lower-ETH minipools get more commission per ETH as a result of matching up their smaller ETH contribution with more protocol ETH
  • Folks that want to reduce RPL exposure can run higher-ETH minipools - Since more people are incentivized to start minipools (due to now being able to buy in, being attracted to more APR, or reducing the perceived risk of RPL exposure), the protocol will be able to mint more rETH
  • Since the low-ETH end has greater projected ROI, and more RPL is used per ETH on that end: it’s expected that this will increase the value of RPL. This wasn’t a key goal (though not damaging value was), but it’s a nice bonus.

A couple of other discussion points:

  • Looking to get RPL to be the first collateral used instead of ETH (and this should get bundled with the above so folks are incentivized to update to this contract)
  • Unclear what the smallest minipool allowed should be. 4+2.8 has been floated, as has 6+2.6. This decision should depend on thoughts around MEV theft.
  • Active discussion around whether 32+0 minipools should be allowed. They could help RP by participating in the smoothing pool. They would want to participate b/c of the smoothing pool and b/c the RP stack is easier to use than setting up a validator without it. However, some folks think we should be looking to serve liquid staking and should therefore require some protocol contribution.

Shoutouts to @Marceau for pushing against letting RPL stake scale down with lower NO ETH, @Mig21 for first bringing up the idea of scaling against protocol ETH, @swingbomb for bringing a similar idea up, @ken for putting up a spreadsheet with some analysis and #trading in general for a great discussion.

23 Likes

TBH the argument that RPL should be the first collateral instead of ETH is a little lost on me - can anyone share more on this?

I support the notion that RPL collateral required for smaller minipools should continue to represent 10% in ETH value of the amount of ETH ‘borrowed’ from the minipool. This lowers overall capital requirement of Node Operators while maintaining almost-equal risk to rETH holders, assuming that the ejection balance is increased as expected.

Thanks @Valdorff! I appreciate you summarizing and posting. Agree on all the points made and am happy to see this idea get raised with broad consensus. :pray: It’s a win for all market participants and also accrues value to RPL. This is exactly the kind of idea we should be pushing for.

3 Likes

So I completely understand the argument for min RPL being calculated as a percentage of the pooled ETH, but can someone help me with the justification that the max should also be based off that?

Staking anything beyond the min RPL is a(n optional) bonus for an operator. Here, we’re no longer considering what’s sufficient protection for the protocol, we’re considering what’s a fair upper bound reward. In that case, is the intent really to allow operators to stake more RPL by staking less ETH? To me, it seems like the cap should be based off the operator’s ETH, not the pool’s. Am I thinking about this incorrectly (@Marceau)?

1 Like

Thanks for bringing this up! I may be put on this earth to express unpopular opinions, but strong vote of no here. I’ll start with an example of trying to attract a solo validator to rocket pool, and then some explanation.
Assumptions for ease: post-merge, validator APR 5%. Node commission 10%

Solo-validator: 32 ETH @ 5%- 1.6 ETH per year

Converting to rocketpool @ 4E: Sell 14.9 ETH, save 25% of that (3.7ETH) for long term capital gains (us centric, sorry), use the other 11.2 ETH on RPL. Start up 4 minipools (16 eTH) + 2.8 each RPL (11.2 ETH). Spend ~0.5 ETH for transactions, node initiation, 4 minipools. Swap 0.6 ETH for rETH.
So now we have 3.7+11.2+16+0.5+0.6 = 32; math works out
This node operator will get 16ETH x 5% x 1.7 (10% node commission x 7) = 1.36
Add in the rETH, another 0.6 x (0.0.5 x 0.85) = 0.026
Total expected yearly earnings for this solo validator to switch to rocketpool: 1.39 ETH (instead of earning 1.6 ETH with no added hassles as a solo validator)

In other words, this proposal doesn’t come close to competing with a solo validator for yield- not even discussing the increased smart contract risks, tax burdens, exposure to another asset, and annoyance of doing the above (and don’t get me started on the difficulty of trying to advertise our new 4 ETH minipool on bankless: “hey that’s great- how much ETH do i need?” “oh just a little under 7… unless you bought your ETH this year, then you are really screwed.”)
You’ll notice I did not count RPL rewards, because of this paradox: if this proposal was in fact capable of bringing in many new entrants, the RPL yield minus the inflation would come very close to zero. It certainly cannot be counted on to stay near its present value.
Essentially, I think this proposal will give existing RPL holders a place to stake their excess tokens, which will sop up some of the liquidity and drive down RPL rewards. However, this will not drive up RPL price unless there is demand from outside the system; to me this proposal wouldn’t convince many people to join (either on the high end or the low end of the economic spectrum).
My belief is that rocket pool is composed of 3 competing interests: NOs, RPL, and rETH; however, rETH generally seems to get the best deal whenever compromises are to be made. It seems already markedly overcollateralized, and this makes it worse.
My recs: keep RPL minimum at 10% of the holder’s stake as long as it meets MEV security requirements (whether that is 4, 6.4, or 8 ETH). This encourages people from outside of the system to buy it (as it will have utility) and will maximize NO’s ETH profits.
Also: continue to push towards non-ethereum bonded pools (RPL collateral), as this will most assuredly help with RPL price and utility.
My 2 cents, sorry about the length. Please send me any corrections, I’m trying to learn.

Won’t the transaction fees make it unprofitable to create a minipool with 4 ETH?

Other than that, I tend to agree with the current consensus.

Topic: Using RPL collateral first
My understanding is to emphasize the role of RPL as primary protection of the protocol’s ETH. I find this slightly clearer myself, but I agree that in the end it doesn’t much matter where protection comes from as long as it’s protected.

1 Like

Topic: What should RPL max stake be based on?
I had the same initial reaction as you about max RPL stake, but then changed my mind.

The thought is roughly that we’re rewarding (A) the NO for allowing the minting of X protocol ETH and (B) the NO for providing a certain amount of protection per unit of protocol ETH. Since this design scales RPL stake with protocol ETH, I think we’re hitting both of those quite well.

Imo, there is an argument to be made that to address (B) we should consider all collateral (RPL + NO ETH), if we did this: right now we allow up to 17.6 ETH worth of RPL to get rewards when used alongside 16 protocol-ETH (a ratio of 33.6 total collateral to 16 protected); using this ratio for a 6 NO ETH minipool, this would imply rewards up to 26*(33.6/16) = 54.6 ETH of total collateral, which is 48.6 ETH of RPL collateral. The simpler 1.5*prococol_ETH design would hit 39 ETH of RPL collateral. This view allows even more RPL in the small pools. I would be against this simply due to additional complexity explaining it.

Keep in mind - additional NO ETH in a minipool doesn’t add value for the protocol, it decreases value for the protocol because we can’t match as much protocol ETH (there is a need for a minimum to avoid MEV theft though).

1 Like

Topic: scenario where this generates more RPL staking, but not more RPL buying
I have three responses here…

1: It would be very surprising to me if RPL could be mass-staked and made less liquid without that increasing its price. Supply and demand implies that reducing the supply should drive up the price. I think the claim is that the additional RPL staked here is RPL that wouldn’t be on the market - regardless of whether it was staked here or held in a wallet. Obviously, I can’t know that, but it doesn’t ring true – I can say if it were available today, I would personally be on the market for more RPL so I could access the more rewarding low-ETH pools.

2: One of the things I really like about this design is that it gives NOs freedom to choose their ETH:RPL balance. If you think the correct valuation for RPL rewards is near-zero, then you could opt for the ETH-maxi route and go 31+.1. Using most of your assumptions, that would hit 31*.05+.1*1*.05+.9*.05*.9=1.5955; unsurprisingly this approaches solo validation, and you get to use the RP stack (and perhaps smoothing pool).

3: In the end, if folks are really bearish on RPL, I don’t think they’re our target audience for potential NOs.

2 minor math quibbles

Neither of these points materially affects your points, hence hiding this in a spoiler.

  • 1 minipools worth of fees could be neglected since solo validation also has fees (that’s why my example above didn’t include fees)
  • The rETH commission should match the NO commission (yours may be immediately more accurate, but long term they’ll converge)

Topic: Fees are more painful for small minipools
Fees could indeed push people away from small minipools. When I set up my minipool, my fees were a bit over .1 ETH. Obviously, that’s much less painful as a percentage of 17.6 ETH than it is as a percentage of 4+2.8=6.8 ETH. Still - even on the 6.8, it’s less than 2% which might be worthwhile if you’re in it long-term. Fees could also be lower as a percentage based on: low gas prices, deciding to stake more than minimum RPL, setting up multiple small pools (some fees scale with number of pools like the smart contracts, but some don’t like buying RPL).

In any case, I think the key here is that 4-ETH pools are an option, not a requirement. NOs can determine the minipool size that makes sense to them (or simply choose to buy rETH if none do).

1 Like

Hi,

Good plan to also give small shareholders a chance to work as a node operator.

Has the rewards of NO (15%) been considered? If I had 40 eth I’d rather make 10 minipools where I grab 10x 15% extra as NO + higher chance of block rewards. Doesn’t this cause problems for the protocol and RETH?

Sweet- didn’t know about this ‘spoiler’ option. this is my response to Valdorff, but i didn’t want to claim a ton of forum space.

Summary
  1. RPL price only will go up if people buy it, not just from removing excess supply. The analogy is if i owned a shoe factory. 1 million shoes in stock, selling 100 per day. One of the warehouses burns, and i now only have 500,000 shoes. Will the shoes sell for twice as much? No- the demand is unaffected, and there is no real supply shortage. We need to be encouraging demand.
    Also- you can buy RPL currently on dexes. Not sure how to interpret your last sentence.

  2. If the best we can hope for is approaching solo validation, why would anyone become a NO? So much risk with neutral or negative reward. The reward needs to be significantly higher than the less risky option.

  3. I am (or was?) bullish on RPL, which is why i’m making these points; i bought a bunch and am running overcollateralized minipools. This proposal makes me bearish because it will not encourage growth and some other competitor will sneak in and eat my lunch.

Math quibbles: my math is correct on the first; this is the math to attract someone from a solo validator (the holy grail of post-merge growth); also solo-staking is much much cheaper than spinning up a minipool. My math on the second is wrong- i was originally using 15% NO commission, but then i changed it after seeing Ken’s chart- but forgot that one.

as pointed out before, with the RPL bond scaling with the “borrowed” ETH, i think a boost on RPL value of some sort should be addressed.
also i propose to leave 15% commission and let the future PDAO vote decide

3 Likes

+1 to delaying commission change. It may still be the right thing to do, but given all the complexity here it makes sense to treat those as separate changes and therefore vote and deploy them separately since there isn’t strictly a dependency between them.

2 Likes

Why wouldn’t it make sense for the collateral range to directly depend on how much ETH is being taken from the protocol? I think you’re making an argument for the collateral maximum to be reduced (as a percentage), irregardless of how much ETH is deposited? If so, IMO that’s a separate decision that doesn’t directly depend on this change (can be handled / voted on separately).

Staking anything beyond the min RPL is a(n optional) bonus for an operator. Here, we’re no longer considering what’s sufficient protection for the protocol, we’re considering what’s a fair upper bound reward.

I don’t think this is strictly true. Reducing NO collateral is ultimately a decision of risk vs. reward. Higher collateral is less risk. Therefore NOs should be rewarded for putting up higher collateral. There are diminishing returns – yes – but IMO it’s not as simple as saying that anything above min. collateral does not benefit the protocol.

Very simply, we shouldn’t design the economic model around taxes, swap fees, or the APR drag for unstaked ETH. I don’t think this way of looking at profitability of NOs by including those factors makes any sense.

The only caveat is that node operators do need to be profitable after a reasonable period of time and they have a higher upfront gas burden. Arguably this is what the commission rate is designed to do.

In other words, this proposal doesn’t come close to competing with a solo validator for yield

This is strictly not true. In fact, @Ken’s spreadsheet proves this: Adjustable Minipools - Google Sheets

To follow your example, a 4+2.8 minipool has a blended APR of 9%, whereas the current 16+1.6 minipool is 6.4% and solo staking is 5.7%.

2 Likes

The spreadsheet is correct (as Ken always seems to be), but the interpretation of data is another thing. I’m not going to hide this because I think it’s important for people to see, cause this seems to be the root of the confusion. I’ll use the 4E minipool as an example.
First- that 9.7% is of only 4 ETH (the other 2.8 is in RPL); so the comparison is 9.7% of 4 (.388 ETH) vs 5.7% of 6.8 (.388 ETH). So yearly ETH rewards are the same for solo validators and minipool operator.
Because when you boil it down, you are telling people- pay 70% more in RPL collateral and we will give you 70% more in rewards. Then tack in all the extra gas costs/tax burden etc/risks etc.

Second- The RPL APR of 7.9% is not counting 5% inflation emissions. so the real APR is actually 2.9%. This is the trap that also confuses people at Olympus Dao and Wonderland (80,000% rewards!). So let’s say 2.9% of 2.8 (.0812 ETH) yearly. That’s your profit over a solo validator, not counting costs. So you might break even with a solo validator in like 1-2 years. That’s if the protocol doesn’t grow; if it grows your effective APR rewards may be zero or negative (after 12 million staked).

What are you suggesting is being hidden? We’re talking about math here.

Yearly ETH rewards are not the same for minipool operators vs. solo validators. Are you omitting the commission? This math does not match reality. Even if you assume the most pessimistic model possible (RPL yield at 0% or RPL trending to $0), minipools are still higher yielding than solo validators. This is because it takes 9% of your position (10% collateral) to accrete +15% commission. This is net-positive in terms of yield generation.

This isn’t wrong, but you’re using really inflammatory language to describe it. Clearly RPL has a different tokenomic and risk/return profile than ETH, and if a particular NO wants to be pure-ETH then Rocket Pool isn’t for them. That isn’t what this post is about.

I do think that the net yield for ETH (likely deflationary) will be greater than the net yield for RPL (currently 5% inflation). This isn’t necessarily a bad thing. I expect that NOs will have an underperforming yield on the RPL side, with an overperforming yield on the ETH side (because of the commission structure, see previous point). This will trend towards an equilibrium and justify a certain RPL price given the cost of RPL ownership vs. the reward of enhanced yield.

RPL has intrinsic value in terms of it’s ability to unlock commission. It also has speculative value in terms of protocol growth. That’s what is fully clear and transparent in the tokenomics and part of what node operators sign up for.

I think @epineph has a valid perspective (with the big assumption that RPL yield is valued at 0). In that case, we should calculate ROI as the ETH return divided by the total investment, and we do in fact get the same result as solo staking if we have a 10% commission and require 10% of protocol ETH as an RPL stake.

Math

ETH return = NO_ETH*stake_apr + Protocol_ETH*stake_apr*commission_pct = stake_apr*(NO_ETH + Protocol_ETH*commision_pct)
Total Investment = NO_ETH + RPL_stake ≥ NO_ETH + Protocol_ETH*min_RPL_pct_of_protocol = NO_ETH + Protocol_ETH*.1

So with minimum RPL staked:
ROI = stake_apr*(NO_ETH + Protocol_ETH*commission_pct)/(NO_ETH + Protocol_ETH*.1)
I tried to put this result in a form that shows the impact of commission clearly. As we can see ROI is exactly the same as a solo staker if commission_pct=.1, which is the case @epineph mathed out (and also what Ken used). @Marceau mentioned a 15% commission, which would indeed give better yields than solo staking.

Now - on to interpretation!

I do agree that if we swapped to 10% commission (not yet decided), then we’d replicate solo staking with more complexity in some ways (more smart contracts, an extra token; but I think more user friendly, fwiw).

Where I disagree is on the perspective on RPL. What I’m seeing here is that I can match ROI on ETH rewards alone, so every bit of RPL rewards is my bonus. In this case, the only reasons to prefer solo staking is (A) I believe RPL APR will be 0, or (B) I believe RPL value will drop with respect to ETH. If you believe either of those… I agree, this is a bad investment and I don’t recommend it.

I don’t believe either A or B. The cases where we have lowest RPL APR are also the cases where the protocol has the most adoption, which bodes well for the RPL:ETH ratio. I believe it’s most likely that the value of staked RPL will increase; this might be due to APR, ratio, or both.

TL; DR
Ken’s take looked at the ROI ETH in vs ETH out and RPL in vs RPL out. This is valid. I also like this way that looks at ROI ETH+RPL in vs ETH out, where RPL out is just… extra. Sounds awesome to me (and ofc even more awesome if commission is >10%).

1 Like

@epineph - if we keep the commission at 15% for now (as suggested by @Mig21 and @Marceau), would that be a clear enough value prop for you?

I think this conversation has gotten extra challenging because there are 2 different commission rates being discussed, and ofc that impacts perspectives.