Yesterday, ICON's social media channels were chattering about an article published in CoinDesk entitled "Everyone’s Worst Fears About EOS Are Proving True," which is a lengthy piece detailing the trials and tribulations of the EOS governance model and the fact it's succumbed to cartel behavior and centralization. This article sparked a healthy reddit discussion on the issue, in which I weighed in on. I've taken those thoughts and re-produced them here with a bit more depth and clarity.


To begin, let's take a look at the CoinDesk article and exactly what the problem is with EOS:

Luzgin left, in short, because he said it is no longer possible to earn funds for maintaining the blockchain without support from major EOS whales, the colloquial term for those with very large token holdings. Those whales are overwhelmingly supporting BPs located in China. There are 21 BPs at any time who establish consensus on the chain, make governance decisions and earn substantial rewards.
This has become a widespread point of concern among longtime participants in the EOS community, for reasons that include centralization and threatened censorship-resistance, according to an investigation by CoinDesk.
The BPs that Luzgin believes have the strongest technical proficiency have overwhelmingly been relegated to lower-tier rewards or no rewards at all.

This has led to productivity issues for the network:

Some sources hint that there have been bad signs for the fundamental performance of the blockchain. Others say everything is mostly fine.
“We see top BPs missing not just blocks but whole rounds,” Luzgin told us. “They are supposed to produce 12 blocks. They miss all 12.”

It has also led to a lack of producing products to help the ecosystem grow:

One of the complaints that members of the community often make about the current crop of BPs is that they don’t prioritize building new dapps that will attract other users to the blockchain.

In short, a collection of whales has been able to concentrate power in the EOS ecosystem in order to reap the rewards of being a "BP," the EOS equivalent of a P-Rep. Rather than using those resources for the benefit of the ecosystem, they're instead primarily turning them into pure profit. Meanwhile, the community is seemingly helpless to stop it.

On first glance, the general description of the EOS governance model sounds familiar to ICON:

DPoS can take many forms, but on EOS, 21 nodes have all the power over the chain. These 21 nodes are chosen by token holders, who stake EOS coins in a vote for up to 30 BPs. The ones with the most votes serve in the top 21, and the vote is effectively continuous. That means BPs can move in and out of the top 21 at almost any time.

This is what has sparked some concerns among the ICON community, with a number of individuals expressing concern that ICON may face the same fate as EOS, due to their similar structures.

Fortunately, there are a number of lessons that ICON has learned from the EOS process that they have incorporated into their governance and IISS models. I will examine a number of areas where ICON has implemented measures to prevent the type of behavior that has plagued EOS. This should not be viewed of as an entirely exhaustive list.

Voting Incentives

Unlike ICON, EOS provides no economic incentive for the community to participate in governance. ICON, meanwhile, provides daily rewards to those who vote for P-Reps. The only incentive provided for EOS voters is the opportunity to participate in governance – for some EOS holders, this may be enough of an enticement; for many, it clearly isn't.

Here is a quote from Ricky Dodds’ research on DPoC (part 1 and part 2)

As a point of reference, there is no economic incentive for voting for an EOS BP. The only benefit is governance. As of now, ~25% of EOS community members have voted, and 47% of EOS outstanding is staked. For Tron, the amount voted for super representatives is even smaller at ~7% based on total supply.

Meanwhile, BP candidates do have an economic incentive to vote, as if they are successful in voting themselves in as BPs, they'll receive lucrative rewards.

The ultimate outcome is that, in general, BPs are participating in governance while the community at large isn't. In an ideal world, the community should serve as a check on the BPs, ensuring more votes are provided to those that provide the best value for the network. Instead, BPs are successfully voting for themselves, regardless of their own performance, as there is limited community power to counterbalance the power of BPs.

Voting Ratios

I'm sure you're familiar with how ICX voting works by now. Each ICX you own is worth one vote. This is fairly simple and straightforward.

Meanwhile, EOS has established a different model. In their case, 1 EOS = 30 votes. However, this is not simply receiving 30 times the voting power. Rather, you're entitled to 30 "slots" that you may vote for a BP; meanwhile, you cannot allocate more than one slot for a given BP. Even this doesn't seem like a big deal, but let's break out the practical impact of this type of model:

Let's say you're an EOS whale with 1,000,000 EOS and you want to vote yourself in as a BP. Under the EOS voting model, you don't have the ability to provide yourself with 30,000,000 votes (even though 1 EOS = 30 votes would imply that on the surface); instead, since each vote allows you to vote for up to 30 BPs (but only each node can receive only one of those votes), it means you can only provide yourself with 1,000,000 votes. However, if you find 20 other BP candidates and band together as a cartel, you can organize to establish vote trading. You can create an agreement among the cartel that ensures each cartel member votes for all other members of the cartel, and only members of the cartel (with EOS, you aren't required to fulfill all 30 slots). If this is successful, then you now have 20,000,000 votes (assuming the rest of the cartel has 1,000,000 votes). Those who aren't organized (or have far fewer EOS) will be stuck with far fewer votes, since they can't rely on the votes of others like the cartels can.

Meanwhile, with ICON, if you have 1,000,000 ICX and you want to vote for others, you're only diluting your own ability to vote for yourself. Each of the other cartel members would be receiving only 50,000 votes of your votes, and you'd be receiving 1,000,000 in total (assuming each cartel member gave you 50,000 votes). This puts you exactly back to where you started: receiving 1,000,000 votes (I am also assuming all the cartel members have 1,000,000 ICX for the sake of simplicity here).

This problem has become so apparent that the CEO of Block.one (the creator of EOS) urged the community to adopt a 1 EOS = 1 vote system. Here is an article going into a bit more detail about that plea.

Rewards accumulating among the community

Over time, and in aggregate, the ICONist community will receive more in rewards than P-Reps. This means, over time, the voting power of the community will grow at a higher rate than that of P-Reps.

As a refresher, ICX rewards are distributed to three groups:

  • P-Reps
  • ICX Stakers
  • EEP / DBP Projects

Although P-Reps as individuals receive significant rewards in order to have resources to help the ecosystem grow, they will–in aggregate–receive less than those who stake, combined with those who earn rewards from EEP / DBP projects.

Based on how IISS is set up, even if P-Rep rewards are relatively high at a given point, and even if staking rewards are relatively low (due to high participation), P-Reps will still earn less than the rest of the community in aggregate.

To better explain, here is how ICX rewards will be distributed under the following conditions:

  • Monthly P-Rep rewards are set at 50,000
  • The price of ICX is $0.40
  • Staking rewards at 15% annual (meaning 30% of the network is staked)

As you can see, on an annual basis, P-Reps (nodes) will receive only 36% of ICX rewards, while the rest of the community will receive 64%. As mentioned, even under extreme conditions that would "favor" P-Reps receiving more rewards, they still remain below 50% in aggregate.

Accordingly, voting power will flow away from P-Reps to the community over time, rather than the other way around.

Vote Buying

Another quote from the CoinDesk story:

There had always been a concern that EOS BPs could protect their lucrative position on the network by sharing block rewards. A top BP can earn about 900 EOS a day, which works out to over $1 million per year at current prices.
...
In an email to CoinDesk, Newdex, a decentralized exchange on EOS, explained how its affiliated company, Newpool, spends its block rewards.
“More than 90 percent of the pool rewards will be given to token holders, that incentivizes token holders to hold EOS tokens for long and also increase their participation in the community,” a spokesperson for the company told CoinDesk.
Another highly rated BP, Big One, also welcomes participants to stake votes for the blockchain and profit for doing so. Infinity Stones invites staking on any number of protocols, including EOS.
“The thing about vote-buying is that at this point there’s technically nothing against the ‘rules’ about it,” Cox said. “There’s not much value-add happening in this new market dynamic, just the shuffling of meaningless inflation.”
Whatever anyone thinks morally about the optics of vote-buying, the consequence is this: BPs engaging in vote-buying invest in ensuring their leadership positions and not in the actual ecosystem. That’s what frustrates those who want to see EOS grow.

If you've been following ICON at all over the past several weeks, you're likely well aware of the debate that's been occurring regarding vote buying. The above paragraph should make it clear as to why so many--including Min Kim himself--have taken such a strong stand on this issue.

Without any stated prohibition to the behavior–and without any way to enforce such a prohibition–EOS vote buying has become prolific. The result has been to ensure the current top BPs are able to further cement their position while also doing little to add value to the ecosystem, since all of their rewards are going into the pockets of their voters.

The complementary role of C-Reps

While all the focus at this point has been on P-Reps, C-Reps (community representatives) will also play an important role in the governance of the ICON network.

Here is how ICON describes the role of C-Reps:

A C-Rep, representing a Community connected to the ICON Network, is elected via autonomous decision-making within the respective Community in accordance to sufficient contribution verified. A C-Rep is a connection between ICON Network and Community, validating transactions on the ICON Network. A C-Rep proposes and votes on policies that represents interests of each Community.

2infiniti explains why this will be a helpful bulwark against the chances of collusion and centralization within ICON:

C-Reps will be well vetted and will be proposing and voting policies of best interest to their respective community. We will have a complementary system among the two sectors of representatives (P-Reps and C-Reps), making our governance structure more balanced and reliable. Also, from ICONstitution and Governance Yellow Paper, we know that the number of C-Reps and the P-Reps will roughly be equal. This makes collusion at least twice as difficult, and prevents biased-decision making.

While the exact specifics on how C-Reps and P-Reps will govern together has not yet been released by ICON, the fundamental power dynamic between the two will play an important role in further decentralizing power within the network.

Conclusion

As you can see, the odds that ICON sees the same issues that EOS has seen are highly unlikely.

I know some have expressed frustration that it has taken ICON as long to decentralize as it has, but as you can see from the above, the ability to learn from the mistakes or failures of other projects can be immensely invaluable, so patience certainly has it's upsides.

Does this mean ICON is completely immune to collusion? No. But no blockchain likely can be. Here again is 2infiniti, on the question of "Will ICON be Collusion-Safe?":

I would say that the answer is no and argue that it is likely impossible for most if not all blockchain protocols to be completely collusion-safe. ICON however will be collusion-safe until a high bound, and fairly collusion-resistant overall.

In general, if you have fears that ICON may become like EOS, you should rest easy.