Currently, there are only payouts for claim assessors if there are votes for approving a claim. Then the assessor can vote to deny or approve the claim to get a payout. However, they still have to put in work assessing the claim to figure out if it should be approved or not. Currently, there’s an incentive for claim assessors to approve claims to get paid; but, no incentive for the initial assessor to outright deny the claim and allow the default deny decision.
There should be some type of yield just for locking tokens up for claim assessment. Ideally, lower than the fees collected for staking tokens on protocols. Anyone open to this discussion?
If a claim isn’t valid, no NXM need be locked up and no vote has to occur. This design eliminates a potential griefing attack, where a user repeatedly submits a claim that then requires assessors to vote, even in high gas environments.
Providing yield for staking NXM in Claim Assessment creates an incentive for members to lock up their tokens without a requirement for work to be performed and, even if yield were provided, that doesn’t require a member to vote, either. Again, if a claim is determined to be invalid, then no vote needs to occur. Providing a base level of yield would create thew wrong incentives, imo.
However, I do recognize that assessors put in work to assess claims regardless of whether or not they are valid claim requests; assessors cannot reach that conclusion without first assessing a claim request. I think the most recent claim submission (Claim V2 24) was an easy claim to assess. In contrast, Claim V2 23 required more work to verify losses onchain. Not every claim requires the same amount of work.
I’d be open to discussing how to create appropriate social incentives to encourage members to review all claims, not just the valid ones. There’s room, imo, for a small allocation of NXM to be earmarked for rewarding members who assess claims that turn out to be invalid. There should be sufficient safeguards to ensure people don’t just spam the forum to claim they reviewed the claim and request reimbursement, too.
From your perspective, @Zarathustra, how much time do you spend assessing the average claim and what do you believe appropriate compensation should be for assessing claims that turn out to be invalid?
Probably half an hour to an hour. Usually, if they’re denied it takes far less time than double checking everything for approving a claim. What the people trying to defraud us are doing is attempting to make complicated claims that confuse the claim assessors. Seems like they’re trying to incentivize assessors to just approve claims.
I think the payout should be half the amount earned from approving a claim. Definitely, less than providing coverage by staking. Something around the current Fed funds rate probably makes sense for just allocating NXM for assessment. Maybe a bit lower. Something around 3% to 4% seems reasonable. Should probably adjust based on the risk free rate.
Using a reduced formula similar to the claim assessment rewards seems prone to being exploited by spammers, and a bad idea. Probably makes more sense to just have a consistent yield that requires registering a claim assessment on chain once per quarter if there are any claims to process. If no claims come in the account just stays registered for the yield. To prevent people from just signing up for free yield, if the account misses a percentage of claims requiring voting per quarter they don’t receive yield.