[PPROPOSAL] [DRAFT] Use MIB's upgrade handler to give Gaia the ability to use minimum commissions

[jacobgadikian]

Regardless of thoughts on minimum commission rates, policies, and the free market, I think that it is quite reasonable to give Gaia the ability to enforce minimum commissions.

feature: implement ante handle for min commission

Is a PR from the MIB validator team that does precisely that.

It should be noted that this upgrade handler won’t change the minimum commission to 5% unless prop 76 passes.

Proposal Text

Vote YES to implement minimum commissions support with either: feature: implement ante handle for min commission by MiB-Intern · Pull Request #1698 · cosmos/gaia · GitHub or similar.

Vote NO to reject the implementation of the upgrade handler / antehandler needed for minimum commissions.

Vote ABSTAIN to express no interest in the matter.

Vote NOWITHVETO to cause depositors to lose the deposit on this proposal and contribute to a failure threshold that is triggered when nwv hits 1/3rd.

[hxrts]

I don’t have a strong opinion but I’d like to highlight that this introduces a potential attack where a cartel of large validators with economies of scale could theoretically raise the minimum commission to starve out smaller players.

I actually agree with the rationale and overall direction of this work, but I also think it’s important that there be some public consideration of adverse effects.

[jacobgadikian]

Hey Sam,

I keep finding this particular concern to be backwards.

EG: in my experience, the minimum commissions don’t starve out new and smaller players, but instead improve the conditions for smaller players.

@jelena did raise a great point on twitter concerning validators with commissions at 100%, though, and 100% commissions have captured more of the VP than 0% commission validators.

Could you please describe how the attack that you think may exist, works?

I’ll model one for you:

The 0% / bribe attack

I am a whaleish with no ethics. I care only of profit. Thus, I choose to make several 0% validators and then distribute NFTs to my delegators and sweep the floor of that nft collection constantly son that my delegators get some kind of fixed value for the NFT’s, from me. But I am also amassing VP. While amassing VP, I am also building short positions under 1000 farmed identities.

Once I reach 33% VP, I halt the chain, and I profit massively on my shorts. If no one figured out what I did, I’ll do this again and again, cackling like Mr. Burns on the simpsons.

[hxrts]

The attack would be the top 7 validators collude to increase the minimum commission to 100% until delegators unbond and small players drop out and people move their delegations to the only remaining validators. Collusion could be very weak, just voting together without any other coordination. Monopoly then allows for price control. I think social consensus will likely kick in before this happens but it is possible.

Btw I completely agree 0% validators are a problem. They are largely marketing operations who are running operations as cheaply as possible.

[jacobgadikian]

Hi,

Gotta disagree about 0% ers.

We should take them seriously, because these “marketing operations” – definitely want to monetize at some point. So they can do a 0% / bribe attack, or they can eventually just raise commissions on their (usually unsuspecting) delegators.

The attack you described is non-viable:

Top 7 is enough to halt the chain, this is true. But it would take 50% of all bonded stake to push through a gov prop to slam the minimum commission to 100%. If we look at the 50% set (66% picured above) we can see that there are many indies in there, whose delegators would override their votes.

Furthermore, we should account for delegator agency & the override. I think that it is much more likely that:

True story about minimum commissions

Notional is doing pretty good now, and our members write all kinds of code in cosmos.

Notional’s seed capital was 1x Uniswap airdrop, sold near the top value of that airdrop. Nothing else. No VC backing, certainly no support on the cosmos hub despite tons of code contributions. We got here because osmosis had minimum commissions and the team and community (both) saw our work as valuable and delegated to us. There was truly nothing more to it.

Notional is helped significantly by minimum commission policies on:

• Juno
• Osmosis
• EVMOS

It’s important for the hub to be able to have minimum commissions, even if that commission isn’t 5%. Otherwise, the most likely outcome here is that the hub should be de-prioritized for validators.

This is an issue, since the hub’s operational complexity is going to go up a great deal with ics.

Race to the bottom

In a race to the bottom scenario, delegations from the icf are meaningless. All delegations are meaningless if they occur on Gaia. We validate the hub solely for brand image. We cannot pay anyone with hub delegations. For us, the hub giveth zero, the ultimate outcome of a race ot the bottom scenario.

Also, if we’re going free markets, there’s a global fee module in the next version of the hub. I think we’d need to remove it. Fees should be zero because free markets.

[lexa]

I think this is worth backing up a step and asking a more theoretical question:

What might happen to the security of the network and the distribution of delegations throughout the validator set if minCommissionRate were to be a parameter on the Hub (i.e., One that can be changed by passing a param change proposal)?

I’d also like to step through various commission rate ideas.

• [minCommissionRate = 0.0] We are currently seeing what happens in the validator set if there is no minimum rate. It has required no collusion to get here, so it’s hard for me to see how how parameterizing the commission rate could make this worse.
  *[minCommissionRate = 1.0]I bet we’d see a lot of people unbond rather than switching validators (Sam and I discussed this in a call earlier today) so I think the risk is less that small validators would be starved and more that the security of the network would suffer as ATOM holders choose not to stake.
• [1.0 > minCommissionRate > 0.0] A minimum rate anywhere between 0% and 100% would hopefully reduce the risk of big, wealthy validators undercutting small teams on their validation services (i.e., Attracting tons of delegators with their 0% commission so that the small validators can’t compete and are forced out, then raising their commission once the competition is gone, a la big box stores and mom-n-pop stores).
  • The closer the minimum rate gets to 100%, the more likely delegators are to decide that staking just isn’t worth it. So I’m kind of curious if there’s also a risk that people could collude to set the rate super high, wait until the total bonded ATOM drops, and then self-delegate a ton and end up with more power than they previously had?

Some other questions I pose to the class (without having any answers myself, lol)

1. What other risks are present as the commission rate moves around, and could move around every voting period of people were really motivated to make it happen?
2. How does the interaction of the unbonding period and voting period influence the likelihood of people colluding the encourage ATOM holders to unbond and be unable to stop before a malicious proposal goes through?
3. I don’t fully grasp the detail of how the upgrade handler works - does it make minCommissionRate into a parameter that could be changed by subsequent proposals?
4. Mechanically, how would the commission rate be changed in the future if this tool were to be added to the Hub?