You are viewing a single comment's thread from:

RE: How Vote Incentivization Monopolizes Delegated Proof of Stake

in #dpos5 years ago

It's a general problem of PoS I would say. In PoS you yourself will use your stake to guarantee the best possible outcome in terms of inflation. While in DPoS you would vote for representatives which will make sure of that.
I think that is one of the great weaknesses of PoS and one of the big points for PoW.

Sort:  

I think that is one of the great weaknesses of PoS and one of the big points for PoW.

Not really, the difference between POS and POW is that in the later the block producers are elected by those that can contribute more hashing power as opposed to having more stake. The fact that only a handful of mining pools are effectively able to produce blocks should be enough evidence that POW also leads to a monopoly (technically an oligopoly).

In POW each individual miner has to decide to which mining pool they should direct their hashing power. The rational thing to do is to choose the one that gives them a better chance of getting a higher portion of the block rewards. It turns out that the best choice is to join a pool that already has a high percentage of the block production.

Oh, I definitely agree, I didn't mean to say that PoW doesn't have any problems of centralization, but it has no problem of monopolization like PoS has. We got mining pools which can be a short term threat for the chain, but if people leave the mining pools again it loses the power. While in PoS that's more difficult. Some PoS chains solve this by reducing the stake of the consensus node on "bad behaviour".

Interestingly enough, its not a problem with PoS! PoS certainly has other problems (e.g., many fail to have effective consensus) but the monopolized entrenchment of leadership isn't fundamental. This issue applies to elected positions.

Consider the scenario in of PoS whereby all participants are actively staking. The inflation is thus being distributed proportionately to everyone -- and the effective inflation is zero. No one gains or loses over each other. When a PoS participant doesn't stake, they are forgoing their sort-of "right to protection against inflation", and this value is transferred to those that are staking.

You are certainly correct though that this is a pain point that PoW doesn't have -- because consensus is not determined from within the system, it is determined from outside the system.

Yeah, but things like requiring a minimum stake to effectively be able to mine a block within a reasonable time makes this much more difficult though, doesn't it? Effectively the biggest stakeholders should be able to grow their stake quicker if they auto stake their producing rewards.

No, that's not correct. In traditional PoS, your expected value of staking is directly proportional to your staking amount. The effect of compounding sounds like an issue -- but what you find is that with a probabilistic system the smaller stakers still, on average, approach the same rate of compounding as well.
So long as a large staker and small staker are always staking, their rewards are proportional. There's a bit of statistics involved if you want to prove it, but it is true.

In traditional PoS, your expected value of staking is directly proportional to your staking amount

No, the expected value is your rewards (proportional to stake) minus costs of running a node (generally independent of stake, and in a heavily used system may be quite large). So smaller stakers still lose out. (Or, more likely in practice, smaller stake holders are unlikely to stake at all.)

Your expected value of staking is directly proportional to your staking amount. I mean here revenue.
You are referring to expected gross rewards, and yes, in a realistic environment the overheads are flat per person regardless of stake amount -- but it's hard to quantify this absolutely though (for example, an individual may have the resources to do the staking already and can overlap it since the resource requirements are pretty low for most PoS systems).

Sure, individual situations vary but you can average across small stakeholders and conclude that they lose out as a group. Unless of course the group is self-selected to include only those with the necessary resources, but this is still going to be wealth-concentrating in general.

Another factor which others were trying to convey is the risk-adjusted return per stake is lower for smaller stake even if pure expected value per stake is the same.

resource requirements are pretty low for most PoS systems

I think this basically assumes the system is a failure, no (or at least, if successful only in a very niche way)? For something with "high" usage (say as high as current Ethereum or potentially much higher), the resource requirements will not be low.

Yes, these are fair points.
I think I'm trying to explain that these centralization considerations for PoS, are a bit different than the considerations for elected positions (and the potential gamification of those positions) for DPoS.
In PoS, you could at least expend system-external resources to protect your stake, but in a monopolized DPoS this could become impossible.

Yes I agree, although 51%-type attack scenarios still apply to PoS where enough stake (either with one owner or a cartel) could censor minority stakeholders or somewhat less (?) maliciously charge them to not censor, which amounts to very similar form of enforced monopolization where even external resources can't help.