I wanted to be a Node Operator (NO) at Rocket Pool (RP), I did my research and decided not to proceed. In this post, I will explain my reasoning. I will state the end conclusions and then explain in detail how I’ve reached them while giving a quick introduction to the protocol.
Conclusions: I believe the tokenomics behind Rocket Pool’s token (RPL) to be fundamentally flawed for two reasons:
Its tokenomics can lead to self-fulfilling prophecies. If a NO leaves the protocol he will cash out his RPL and decrease its value. So if a NO fears that others will leave, all he can do to stop his wealth (RPL) from decreasing, is to leave first. This creates a race. Which makes the token value more susceptible to competitors pressure and to market cycles.
Forcing a NO to use an unproductive token like RPL as collateral greatly reduces the protocol yield. This leaves a good margin for competitors to do better and poach NOs.
First things first, what is Rocket Pool and what are its tokenomics? If you already know this, please skip this section.
Rocket Pool (RP) is Liquid Staking protocol. It works as a marketplace, on one side we have the depositors. People who want to earn a safe yield on their ETH but don’t want to handle the hassle (or don’t have the funds) to set up a validator node. On the other side, we have node operators (NO) that stake the depositors’ ETH in exchange for a commission on the rewards (14% in this case).
When the depositors deposit ETH they get rETH in return. rETH value keeps increasing as the deposit earns more and more rewards.
So far, so good. Super clean 🧼
Now it gets messy.
To become a NO, you need to deposit 8 ETH plus 2.4 ETH worth of RPL (the RP’s token). In return, the protocol lends you 24 ETH from the depositors so that you can create a validator node which requires 32 ETH (8 + 24). The RPL is used as collateral to be slashed if you misbehave.
So the demand for RPL increases if more people want to become NO, and the supply of RPL increases if NOs want to leave the protocol.
Additionally, RPL is inflated 5% per year and from that inflation 70% goes back to NOs and 30% goes to pay expenses: like development, maintenance and oracles.
If this seems like a super convoluted way to pay expenses, it’s because it is. I’ve heard that they did it this way to pay for expenses and also to make RPL investors (there was an ICO) happy. But, I still don’t get this reason. I think it would had been cleaner to simply give a percentage of the protocol profits to RPL holders.
Ok these were the basics! Let’s dive deeper in the two problems mentioned at the very beginning:
This text is getting boring so let’s make an analogy with a marketplace we know well. Uber. On one hand we have drivers, on the other we have users that want to be driven.
To become a Uber driver you need to pay for the car and also give us a collateral in case you have accidents. But, there is a catch. The value of this collateral will increase if more Uber drivers join and it will decrease if Uber drivers leave. Even better. It will also depend on what people speculate about the future growth of the number of drivers! Let’s call this idea degenomics.
In the beginning, degenomics is a success. Uber is a fantastic idea. It’s a first mover, has no competitors and its attracting many drivers. The collateral of every driver is increasing fast, which attracts even more drivers and speculators. However, with success come the copy-cats.
Enters Lyft a similar marketplace that offers slightly less commissions to users and higher payments (in stable dollars) to the drivers. It also chose to use economics instead of the degenomics.
Uber drivers had success on Uber and don’t want to move, but they fear the growth is not sustainable and their colleagues might be enticed by the safer and less volatile payment in dollars. If their colleagues leave, their now huge collateral will lose value. It’s best to call it a day, secure the bag and leave before others do. Once this starts, the other drivers see their wealth decreasing. The fear increases. The last one to leave holds the bag.
In sum:
in the real world you get your collateral back if you don’t misbehave.
in degenomics depending on how you time your entry and your exit the collateral you get back is bigger or lower. Ideally, you deposit the collateral before everyone else and you withdraw it before everyone else (including speculators). Basically, a zero-sum game where your performance is given by your timing to enter and exit the game. A ponzi game for short.
Not cool, this is FUD! Rocket Pool is the best protocol out there. There is no Lyft. Is it even possible to be better?
By forcing NOs to use an unproductive token (a token that doesn’t generate yield) the protocol’s efficiency is greatly reduced and there is a big margin to do better. Let’s do the math and figure this out.
We’re going to look at the yield of a RP node operator when the protocol has reached maturity: the number of NO is stable. It’s important to look at this stage, because 1) if it all goes well we will reach this stage, 2) if the protocol is not competitive in this stage then NOs will leave, making the protocol even less competitive.
Assuming a solo validator gets Y% of yield, and assuming RPL is no longer inflating at 5% per year (some say the end goal is to make inflation go to 0 - we’re assuming this is possible).
At the maturity stage the yield of a Rocket Pool NO is: 1.0923 x Y% (9.23% better than staking alone). Math is in this spreadsheet and explained here.
How much better can a competitor be?
Let’s say, the competitor does the following:
applies a 14% commission on the deposit ETH (like RP).
70% of that commission goes to NO, the remaining 30% goes to pay expenses.
Instead of using RPL as collateral, they use rETH (or some equivalent yield earning token).
In this case, the yield of a NO is: 1.1938 * Y% (19.38% better than solo-staking). More than twice as good as RP. The math is in the same spreadsheet.
Note that this collateral is also safer. A collateral in RPL can decrease a lot in relation to ETH, and the NO can choose not to top it up. This means we get a NO running without collateral and no way to punish him for misbehaving.
Tokenomics aside, I love everything about Rocket Pool. The community is awesome, the documentation is excepcional and I believe they have a good heart and are working towards a noble cause.
However, we would never use degenomics in the real world, and we shouldn’t use them in crypto too. The crypto community should not be afraid to discuss these things.
Also, I could be completely wrong about this. Would love to hear everyone’s opinion/feedback as I am actively trying to change my mind and improve my beliefs.
If there are future plans for the RPL token I think they should be revealed. I think this should have been done before the recent sale of RPL to Coinbase Ventures at an undisclosed amount for an undisclosed valuation (another thing that would never happen outside of crypto with public shares).