Only Solution to Centralization Problem: Increase Delegation Penalty and Reward Rate

I believe the only solution to the centralization issue is two fold: increase the delegation penalty massively, while simultaneously increase the reward rate massively.

As it stands, right now, the problem is that node security and trustworthiness is a self-reinforcing mechanism leading to further and further centralization: I, as a small holder, will simply delegate my stack to the most secure node, regardless of how many tokens are already staked with it; in fact, not only regardless, but largely BECAUSE there are so many tokens already delegated to it, which implies that many others view it as secure and trustworthy, increasing the probability of it actually being so.

So we need to take a look at the average holder. How could we incentivize he/she to run a node? Under what conditions would he/she be essentially forced to do so? In my view, under the conditions that:

  1. hodling, and not staking, would lead to significant dilution and

  2. delegating is heavily penalized.

Under the conditions above, I would be highly incentivized to run my own node, rather than hold and/or delegate to an existing node. The only way to force one is to increase the rewards massively, above 50% per year guaranteed. The second is self-evident: FORCE the delegation fee to be above 50% or so, and maybe even burn some part of the rewards that would normally go to that delegator

One variant of this would be to scale the delegation penalty with the size of the node: if a node has say t tokens under its control, increase the delegation penalty proportionally to t. Ultimately, however, not sure this will work, as larger nodes will just break up the tokens under their control into multiple nodes and game the system

Hi, Edmund:

I like the idea of forced minimum commission fee, which definitely will push more people to consider running validators themselves instead of delegating.

Is there any other blockchains that applied this rule?

Seems this is a clear solution to force more node-running than delegation. Interested to know why or why not other blockchains adopt this.


1 Like

Apparently, Binance is acting as the Godfather here. Its participation is so high that centralization is compromised.

The impact on the image of Harmony and the blockchain is not good. We see every epoch the same 20 Binance powered validators fighting each other for the available slots.

1 Like

I actually don’t know: the only two protocols so far that I’ve seen that actually have implemented delegating to begin with are harmony and fantom. Fantom has 50% rewards per year and forced locking of one’s stake for a period of 6 months, under penalty of losing rewards. But then again they dont actually have made independent nodes running on Opera

The idea of delegation itself seems like it’s resulted in more problems then it’s actually solved. While the original intention may have been good, giving people the chance for low effort participation in the network, the end result is just extreme centralization in the validators.

Some sort of minimum penalty would definitely incentivize people to start running their own nodes. But it would need to be significant. But another thing that would incentivize people to run their own nodes would be to increase rewards. If you increased the inflation rate massively, and made a huge delegation penalty, people would have to either sell or run their own nodes unless they want to get diluted by the inflation.

As it stands right now, things will remain centralized until the network is shocked into action with more extreme policies


You can delegate on pretty much any proof of stake blockchain, look specifically at DPoS (Delegated Proof of Stake) blockchains. Tezos is built around the delegation aspect as well but the validators are known as bakers and they’re completely voluntary. On Harmony the blockchain is setup around these 320 seats to maintain a faster block finality, but Tezos has unlimited baker slots. Anyone with over 8000 Tezzies (called a roll) can become a validator (baker). Anyone with less than 8000 tezzies can choose to delegate to those bakers, however there is a cap on how many tezzies can be in a single validators roll. Once that cap is reached, rewards for anyone else that delegates are null and only those that put in before the cap are still rewarded. Thus, decentralization is natural. No one is willing to delegate to someone for zero reward. They must go to another baker if they want to see staking rewards. However, each baker can choose the fee for delegates similar to Harmony. I understand the desire for speed but if we’re risking decentralization for speed, I believe that is a terrible trade off. Blockchains are just bank ledgers without decentralization. That’s just my two cents.

so what is your proposal to solve the current problem? All i see is a description of how things are

My point was that saying this is the only solution to centralization is wrong.

Thanks for feedback.

The 320 seats situation is only temporary and we are working hard to expand the number of seats in the next few weeks to lower the barrier to become a validator so more individual small validators can participate. Eventually, there will be thousands of validator seats for bidding and the entry threshold will be similar or even lower than 8000 Tezzies.



hey Clayford, what would prevent another validator to create a different baker ? so he can attract more delegator under the same entity (thought different baker)

nothing. clayford didn’t think his response through. so far, no one has rebutted my argument convincingly IMO

I quite convincingly proved that your argument is completely invalid. This is not the only solution that you have presented. Not to mention, it’s a terrible solution.

8000 tezzies prevents a validator from making another baker. But Tezos doesn’t mind having more bakers. Someone could essentially have 10 bakers, sure. That would be incredibly expensive not only for the startup costs but also for the node expenses.

no, you didn’t. what stops larger validators from just splitting into multiple validators under your proposed solution?

Nothing. That’s what I want to see. My proposal is to bring more shards online and open as many seats as possible. Along with that, I propose we reduce the amount of BLS keys that nodes can have to 4. I propose we set a min and max amount of ONE that each validator can have to receive rewards (per BLS key). The min will be set there to make the cost of entry higher than the 10000 ONE, (which is nothing for node operators). The max will be there to force validators to make other nodes. This will cause the median staked to go down and allow more nodes to come online. Thus decentralization is born. You can make the argument that node operators can just make more but at that point they’ll have to compete with smaller operators. It actually gives smaller nodes a fighting chance.

i agree, combining with the original point, i think fixing the total amount of rewards a single validator earns would be big, or at least putting a cap on it. something along those lines…but i don’t know if the team has the will to make such a large change so late in the game

but its better to make the hard changes NOW and EARLY, rather than later

1 Like

Hey look at that. We’re coming to a consensus.

us two are. unfortunately, we are mostly irrelevant

also I still believe my original suggestions should be included, but I know they won’t, so I’m gonna stop mentioning them as much

hey @rongjian @edmund thanks for bringing up this topic.
me and @ChainodeTech Ionut had a discussion yesterday on how we can increase profitability for validators and we think a forced minimum commission fee is feasible :] some other projects who have implemented similar rule:

there are protocols where everyone has the same fee, in icon we cannot control the fee, I know in cardano you can set a fixed fee and a variable fee.

fixed + variable fee looks like an interesting idea.