Yet Another Idea on Decentralizing Stake Delegation

I know IOG is not interested in tweaking the current reward sharing scheme, but I just would like to share an idea that I or anyone may explore in the future after Voltaire.

There is a mechanism I am trying to formulate that may help decentralization. It is based on the fact that wallets owned by retails hold more than just a bag of ADA. They hold a variety of coins, whereas private SPOs and CEXs primarily hold ADA.

So, the protocol that I am thinking about calculates how decentralized the SPO is - i.e., how many unique wallets are staked and how diversified (based on unique coins held) these wallets are. The higher the decentralization of the SPO, the more power the staked ADA holds. The coins other than ADA must meet a certain criteria and must be voted on by a governance mechanism.

So for example, two SPOs with the same stake but one is purely ADA and only a few number of unique addresses vs one that includes a variety of approved coins plus more unique addresses. The latter will have a higher decentralization score and, therefore, gets more blocks and rewards.

In this manner, delegators will seek out SPOs that are more decentralized, i.e., delegators will seek SPOs with more noncustodial retail stakers. As you can tell, CEXs and private pools will have lower decentralization score.

1 Like

An interesting idea. I’m not sure I like the idea of earning more rewards based on diversified wallet holdings. But I do like the idea of a decentralization metric on the number of unique wallets in the pool and diversified holdings. Could be called a Public Holdings Decentralization Metric or something of the sorts…

1 Like

Hi @ccgarant, thanks for dropping by. Rewards will not actually increase in this proposal but will be proportional to the pool’s Public Holdings Decentralization Metric, among other things including pool pledge and saturation level.

We might be able to use IOG’s proposed decentralization index. So, this proposal is not going to start rolling until CIP-1694 and the decentralization index are implemented.

Anyone can create any number of wallets they like. Thousands of wallets can be managed by one entity. If you game-ify metrics like this then people will build software to automate the management of thousands of wallets.

The majority of Ada holders are not going to vote for a change to the Ada staking model such that holding token “X” gets you more staking rewards. Furthermore, such a change would undermine the security of Cardano’s proof-of-stake model because it would provide methods for attackers to get disproportionate control over block production.

To quote from the Ouroboros Praos paper:

At any moment, the probability of being permitted to issue a block is proportional to the relative stake a player has in the system, as reported by the blockchain itself.

And then this:

A particularly challenging design aspect is that the above probabilistic mechanism should be designed so that the adversary cannot bias it to its advantage.

Anyone can create any number of wallets they like. Thousands of wallets can be managed by one entity. If you game-ify metrics like this then people will build software to automate the management of thousands of wallets.

Ah, yes. I am quite aware. I am thinking of a number of ways to do mitigate this. As of now, one solution is to make the decentralization metric proportional to the wallet holding. So, a wallet with a zero holding will have no effect on the metric, a wallet with 100 ADA will have a minor effect, while a wallet with 10K ADA will have a significant effect on the metric.

The majority of Ada holders are not going to vote for a change to the Ada staking model such that holding token “X” gets you more staking rewards. Furthermore, such a change would undermine the security of Cardano’s proof-of-stake model because it would provide methods for attackers to get disproportionate control over block production.

None has been written yet. All, I can do is present data to prove/disprove the viability of such mechanism.

The tokens shall pass a set of standards before they can be included in the metric. They are included and dropped from the metric just like how big corporations are included and dropped from S&P500.

I don’t have a proof yet, but I can already see that attackers need more than just 51% of stake ADA to hi-jack the blockchain. As of the moment, all we will have here are speculations unless I have proofs.

As soon as the optimal metric is known then anyone can just reconfigure their wallets and create new ones to achieve the best metric.

The Ouroboros Praos paper is very clear about the security of Cardano being based upon the distribution of Ada where every Ada is worth the exact same as every other Ada irrespective of how the Ada is grouped in wallets. What matters is how much total Ada is staked with each stake pool. Attempting to change this to include weightings for other token holdings, or some sort of variable weighting depending on individual wallet size, would undermine the security of Cardano.

I run a small pool and I am also frustrated with the current reward sharing scheme for several unfairness reasons, but attempting to change it to be based upon some other tokens is… I suggest joining up to the Matrix “Cardano Professional Society” channels where @Michael.Liesenfelt and others are actively discussing the future Input Endorsers (CIP-79) reward sharing scheme design and implementation. That might be the best way to combine efforts to influence the future direction.

You can also have a read of the discussion in the Github repository for CIP-79: CIP-0079? | Implement Ouroboros Leios to increase Cardano throughput by dcoutts · Pull Request #379 · cardano-foundation/CIPs · GitHub

Thank you for your input. ATM there is really nothing to criticize nor defend :grinning: I will provide the data in the future, and that’s when we can decide via governance.

Just a minor input: the optimal metric is high pledge and more decentralized public holding. This is something that I am aiming.

If you look at my pool, I have very high pledge because I have corralled Ada from everyone I know who will trust me, and put it all as pledge. Still on my website I say the following:

I do not recommend you stake with Terminada pool yet because the protocol does not allow setting the fixed fee any lower than 340 Ada.

With a fixed fee of 340 Ada you will lose too much of your rewards to fees until the pool size is over 10 million Ada.

It is very frustrating to me that high pledge has almost no impact, so I very much agree with you aiming to somehow make pledge more relevant.

Another important factor for proper decentralisation is that more physically remote pools are at a disadvantage if they suffer greater network delay. This is simply because they will get their blocks involved in more “fork battles” that are decided by the VRF which produces a 50/50 chance of losing.

Here is a quote from the CIP I am working on:

This might seem like a minor problem, but the effect is quite significant. For example, a stake pool in Australia suffering from only 1 second block propagation delays will receive 3.46% less rewards compared with a pool located in Europe or USA. The numbers are even worse for a pool suffering 2 second propagation delays because it will receive 6.53% less in rewards. Keep in mind that it is not just a reduction in rewards, but also the same magnitude reduction in influence over the ledger, which might be more valuable, and important.

Considering that most stake pools are competing for 1% or less in fees, these are big numbers. The obvious solution is to move the block producer to a server housed in Europe or USA. This illustrates not only the centralisation problem created, but also the reduction in security that follows running a block producer on someone else’s computing hardware.

Come join some other people thinking and talking about such things at Cardano Professional Society https://matrix.to/#/#Cardano_Professional_Society:matrix.org
From the main link all subspaces and rooms can be located.

1 Like

Yeah, I don’t see the point of minfee if pledge benefit is working as intended. Will join in one of those spaces when I have the chance. Thanks!

1 Like