DeFi Passport Issuance - Initial Rules

Background

The purpose of this request for feedback is to set the initial rules for the ARCx protocol and the underlying machine learning algorithm. In essence, we need the community to come to consensus on what identities can, or cannot receive a DeFi Passport and for what reasons.

To achieve this, we must agree upon a reasonable definition for what the community considers a valuable identity to provide reputation services for. It is pertinent to note that these rules are dynamic and can be subject to change as the DAO grows, evolves and gets feedback from the community and open market.

The rules that will initially govern the system are split into three distinct categories, which are as follows:

1. Screening

This will determine whether an on-chain identity is worthy of being indexed. For an identity to be successfully screened, we believe that the following requirements are relevant:

  • At least 45 days of wallet history
  • Has interacted with less than 10 unique addresses
  • Spent $50 worth of ETH on gas
  • Maximum number of tx’s done within a time interval
  • Activity on-chain (i.e cannot be active for 24 hours)

The requirements above are not yet defined and we are hoping that the community would be able to provide feedback on this.

2. Application

If your identity is successfully screened, you will now need to apply for a DeFi Passport. The First Edition Defi Passport will require that an identity stakes 2,500 ARCx tokens for 1 year.

All Passport editions beyond the initial release will require a lower amount of ARCx tokens to be staked for the same duration. Keep in mind that First Edition passport holders will accrue additional benefits as early citizens of the ARCx protocol.

Feedback from the community about staking requirements to apply for a DeFi Passport are welcomed.

3. Slashing

Slashing of passports is only applicable when users cause direct, measurable harm to protocols such as not repaying under-collateralized debt. ARCx will not slash passports for arbitrary or subjective reasons to do with token ownership, selling behaviours or political agendas. We strive to be the credibly neutral reputation layer that the decentralised ecosystem can rely on.

Any feedback from the community on these preliminary parameters would be greatly appreciated.

Thanks all!

3 Likes

For Screening, is there a description of what type of identity should be included or excluded and why? The list is measurable behavior, but is there a summary profile for a good or bad passport holder to target? Other than “no bots”? Because a bot could actually easily be tuned to meet published criteria.

Hi good stuff. Could you comment as to why you would include

Has interacted with less than 10 unique addresses

Do you want to limit the exposure per wallet address to a limited number of DeFi projects? Thus increasing the amount of DeFi passports needed by individuals this way?

2 Likes

Would be interested in knowing whether there is some qualitative/captcha process we could incorporate that still keeps passport application and ownership trustless/permissionless.

Feels like 45 days is too short to be considered. I don’t know what the time frame would be but feels like at least 90 days before someone could be considered.

1 Like

If the goal is to exclude 100% automation, this makes sense. If the value of a passport is high enough, you could see people automate all the stats of the wallet and then manually complete the last captcha step. In the systems I’ve run, the majority of abusive accounts are a hybrid of automation for the things that take time, and manual updates for the critical human check steps.

Hey @Hendisco @pepite @Arcteryxsaurus @acedabook, thanks for all the great feedback! So I think a few things worth clarifying here that may provide more context/clarity:

  1. Initially we were thinking of this being a bot/not-bot measure but given the subjectivity of it, we decided to remove it as core requirement to the protocol. Instead, these requirements are simply to determine which wallets/identities have enough history to start indexing effectively.

  2. The actual behaviour being measured will come down to the actual relevant score, not really whether they have a passport. ie your credit score, yield farming score will all be different.

Hope that provides more clarity and thank you for chiming in here!

1 Like

That makes a lot of sense. Totally in agreement.

I don’t quite get it. Does it mean if my wallet interacted with more than 10 unique addresses I’m not eligible?

Agree, that makes sense. My only other 2 cents would be to think about time for the measurements - partly because defi 9 months age was a much smaller space, and partly thinking about how user behavior must have changed dramatically even between Feb-May to May-July with the market correction. Some of the metrics like 45 days of wallet history could hold, but the Max transactions or the activity on chain could be date bound.

1 Like

I think this was meant to say that less than 10 won’t be eligible :laughing: Defs doesn’t make sense the other way around!

Hm good point, any suggestions to what numbers you might recommend given the distinct period where we say “interesting” market action?

In absence of a finely tuned ML model, you’d want some baseline number for any metric that changes during “interesting” periods up or down, then have like a 1.25X multiple for manic periods (subjectively defined as big runups in either BTC/ETH or big runups in tokens issued), and then a 0.8x multiple for quiet periods. You also want to specify transactions that you count, like swaps (2x/mo baseline average?) or approvals (5x/mo baseline average?).

Something like a wallet accepting an ENS address would only happen once ever, but would be a strong indicator of a valuable identity.

1 Like

When will the passport launch?
The date. I want to know.