I agree with most of your post. But I had to cut out a bit to correct this statement:
If you read the Ouroboros papers, this is exactly what the design is seeking to achieve. The design intends for the reward mechanism to incentivise the stake in the system to shift in order to achieve this target number (K) of pools which are evenly saturated.
In other words, the K parameter has two objectives:
- To ensure there is at least this minimum number of pools. IE: minimum number of pools >= K.
- To push towards this being the optimal number of pools. IE: maximum pools is not too much more than K (with all being mostly saturated).
However there were a few problems the architects didn’t consider:
-
The designers did not envision, and did not model for, multi-pool operators. Multi-pool operation bypasses the minimum policy objective of what the K parameter was seeking to achieve in terms of unique individual operators. With multi-pool operation you can have only a few, or even just one individual operating all K pools.
-
The designer assumed that stake would remain active. The security of the system is based upon active participation where stakers will move their stake if a pool operates badly or changes its fees. They didn’t think about the problem of “sticky stake” which provides a windfall of extra income for OG incumbents, and provides them a “get out of jail free card” since they can act maliciously with less fear of reprisal.
-
They envisioned a world where new pool operators could fairly compete with incumbents. However, as you point out, the minPoolFee is a barrier to fair competition enforced by the protocol. New entrants are not on equal footing with incumbents due to this mandated minPoolFee, and this economic disadvantage is further exacerbated by the OG pools having tons of sticky stake that can’t, and won’t, move no matter how they set their fees.
If they are a multi-pool operator they have even more power because they can spin up additional pools with lower fees and leave their OG pools with higher fees to cross-subsidise expenses. The sticky stake can’t and won’t move, so the operator can just charge high fees on the old pool and low fees on the new pool. Their new pool can out-compete your new pool with lower fees but it doesn’t hurt them economically nearly as much as it hurts you.
But it gets worse: The multi-pool operator can even go on social media and say that they are the good guy here because they are trying to retire their old pool and this is why they are charging higher fees on it. They are “trying to get stake to move and be more balanced”. Have you noticed that some multi-pool operators charge different fees on different pools?
For those single pool operators interested in fair competition here are some changes that would help:
-
Un-recognise sticky stake after a certain period of time - perhaps 12 months.
IE: Change the delegation mechanism so that each delegation has an expiry of 12 months. After that time, users will need to re-delegate or re-confirm their current delegation. IE: Prove they are still alive and can control their keys. After all, the security of Cardano is based upon an assumption that stakers are active participants and they are partnering with their stake pool operator. And, stakers get paid rewards for their ACTIVE participation. Sticky stake is doing no such thing and should not be rewarded when it is not active. Furthermore, if we stopped paying rewards to the inactive participants then this would result in more rewards being available to be paid to the active participants. I don’t think it is too much to expect a confirm delegation transaction once a year to prove you are still alive and are able to point your stake key somewhere.
-
Give stake pool operators the option of choosing a minPoolPercentage instead of a fixed minPoolFee.
At least that would level the playing field for new entrants in terms of not disincentivising stakers due to reduced rewards.
But note: This would in effect disable one of the policy objectives of the K parameter - that of seeking to push Cardano towards having K saturated pools through disincentivising tiny pools.
But then again so does the advent of Multi-pool operation bypass policy objectives of K. So, quid pro quo.