A way towards Scalable and Trusted Dispute Resolution Network on EOS.

EOS network is stepping towards Referendum and debates around constitution are sparking up. Since our last post on introducing Tungsten, ECAF check and balances have been revised by Ian Grigg (Advisor on Arbitration Working Group, EOS Alliance). One of the major modification is — now BPs retain the power to judge the awards and execute based on the rulings.

Previously proposed Check and Balances by Ian Grigg

Current Proposal of Check and balances by Ian Grigg on EOSGov — 
1. transparency of the process — filing, case. Albeit the Arbitrator has some latitude to close off some elements here.
2. the public ruling which some few will make a process of reading and critiquing, which critique will stick to the name of the Arbitrator.
3. the in-forum appeal process.
4. the decision of parties, BPs and other impacted persons to enforce or deny the ruling.
5. the ability of ECAF to provide some non-case pressure. Mostly, no more cases.
6. disciplinary hearings.
7. the impact of the Community referenda to speak as a third head of power.
8. appeals to national courts.

ECAF rulings are now proposed to be non-binding i.e they may or may not be enforced by BPs. Though non-binding award/rulings help with maintaining check and balances, the approach seems to be impractical if one consider scaling and efficiency. Our arguments against the proposal are—

1. Inefficient Structure.

In case of any dispute, a claimant files a claim on ECAF with piece of evidences. If dispute needs an account being frozen (in most cases), an emergency arbitrator gets assigned who looks into evidence (could even demand more pieces of evidence) and give an arbitral order to freeze accounts.

Case details and execution history — https://eoscorearbitration.io/case-ecaf00000441/

Active Block Producers need to judge an arbitration order (as they are not-binding and are at discretion of BPs), reach consensus (21/21 BPs) on whether to blacklist/freeze an account or not. Once consensus is reached, the claimed account gets frozen by BPs.

In simple terms, a claimant needs to act within the 3-days of staking period, collect enough evidence to prove his claim and remains on the mercy of Emergency arbitrator plus 21 Active BPs to work effectively. There already have been two cases where hacker accounts made their way out of blacklist as BPs weren’t able to reach consensus or failed to update the blacklist.

2. Scalability

Expansion of EOS Network will lead to many disputes which will demand fast actions by Arbitrators as well as the Block Producers. We don’t feel the Dispute Resolution structure in its current form could work or effectively scale, putting our arguments forward -

a. Assumption voters will vote out BPs if they don’t comply with legit orders.

Let’s look at the problem of DPOS. On paper, BPs are ‘decentralised’. We can even measure it at 21. And the community will praise and punish via voting.
Except, that’s not what is happening. The community are NOT praising and punishing BPs. What is happening is that a small group of cooperating whales are pushing the BPs up and down according to some internal agenda that they are not revealing. — Ian Grigg (@suntzu)

Whales may or may not care much about the arbitral award by ECAF or the judgement of whales. Even if they do, it would be very difficult for the voters to keep themselves updated on actions of Block Producers whether they comply with an arbitral award or not. Secondly, BPs judgement would be biased towards whales and big proxies to keep themselves in business.

b. Small size disputes

We need to think of a structure which considers $20–100 disputes and is automated. We could not expect human intervention in every possible claim on the network. Human intervention would increase the prices of dispute resolution, in current scenarios $20 disputes may not even be considered. To suggest that BPs should look into every $50-$100 claims will be wrong in itself and has a weak standing.

Any mediation through current model need to be followed by emergency actions by BPs (account freeze or transaction reversed), which should be automated to no human intervention.

c. Inconvenient

Further we believe it won’t be right to assume ECAF and BPs will act fast and safeguard from hacks/scams/stolen funds, it would lead to more disputes and chaos when scammer could transfer tokens to DEX and exchange them for new tokens of TLOS/Worbli.

We are stepping towards mass adoption (with thousand of dApplication and millions of users), it’s too much of workload for a single entity in managing or reviewing external forums/court rulings.

Could we have many forums resolving disputes?

The current structure gives you the freedom of opt-in for different forums at dApplication level but any award by external court needs to reviewed by ECAF to get it enforced by BPs.

Reasons being-

  1. It would be difficult for BPs to consider ruling from many forums opted in by the community.
  2. Any arbitration forum without skin in the game or anything at stake could be influenced either by competitor dApplication/ bad actor in his favor.
  3. Interestingly one point which has been raised many times in Gov chat is if national courts are willing to get their rulings enforced against certain party in their jurisdiction they need to claim on ECAF as Claimant.
Thomas Cox explaining why we need ECAF on the base layer? Starting at 6 min mark.

Need for Fourth Party (Automation) in Dispute Resolution

Though EOS is still in its infancy, it is scaling rampantly and has become one of the most used Blockchains for now. With scaling of the network, automation becomes important in terms of Online Dispute Resolution to establish trust among the community members and offer them hassle free services. Ebay has been at the forefront of online dispute resolution and there are many lessons we could take from Colin Rule (Former Director of ODR on Ebay).

Key points from Colin Rule’s published paper “What we know and need to know about online dispute Resolution” and why we think automation will be needed?

1. The number of disputes increases whenever transactions and relationships increase.
2. The more novel the activity, the greater the likelihood of disputes. The first iteration of an innovative product or activity rarely anticipates all the disputes that it will generate.
3. The more valuable the item or issue in question, the more likely it is that a problem or grievance will turn into a dispute.
4. Speed and time pressures lead to disputes. If the value is likely to erode quickly, as is often the case with technology, a pressure to protect and aggressively extend its value increases.
5. Increased complexity in relationships and systems create more opportunities for disputes. In the words of computer scientist Peter Neumann, “Complex systems break in complex ways.” When informing shareholders about a federal investigation of problems in correcting errors, Experian stated that “We might fail to comply with international, federal, regional, provincial, state or other jurisdictional regulations, due to their complexity, frequent changes or inconsistent application and interpretation.”
6. The easier it is to complain (by filling out an online form or sending an email), the more disputes there will be.
7. The lack of transparency in algorithms leads to disputes.
8. The less attention is given to preventing disputes, the more disputes there will be.

Taking steps towards including Forth Party in Dispute Resolution

Ethan Katsh and Janet Riftkin’s book Online Dispute Resolution had introduced the concept of the fourth party, in which technology could play a dispute resolution role separate from the human third party. eBay following the concept decided to write a software program to assist the parties in resolving disputes and to involve neutrals only on an exceptional basis.

Much like EOS, eBay had a large influx of claims from many countries involved, the disputes were resolved faster with the help of Paypal as a partner following chargebacks mechanisms (A chargeback, also referred to as a payment dispute, occurs when a cardholder questions a transaction and asks their card-issuing bank to reverse it.)

Chargebacks act like the collaterals/escrow which transfers money back (without any evidence) in the customer's account. The vendors have the burden of proving that the merchandise or service was given according to the contract terms, once that is proven the bank transfer effective payment to the vendor.

Further Ebay took the staircase approach, beginning with problem diagnosis and working with the complainant, then using technology to assist negotiation and finally moving to an evaluation phase where eBay and its payment provider, PayPal, would decide the case if the transaction partners could not do so.

The complete structure brings trust, convenience, and technical expertise.

How could we have similar trust and convenience for regular users on EOS or a decentralized network without involving BPs?

In a decentralized network, it becomes difficult to implement chargebacks as one would need intervention by BPs to revert the transactions. The best way around the problem is trusting parties putting the bond or implementing transaction bonds on dApp level.

EOSMetal is working on introducing Bond Structure on EOS Network. Please read Tungsten introductory post to have more understanding around Bond/collateral here.

Bond/Collateral will help with trust and automation (staircase approach followed by eBay). Any claimant need not be at the mercy of BPs or emergency Arbitrators taking the action or use third party Escorws, he/she could claim and put forwards pieces of evidence in the time-period mentioned in the collateral.

Free Market Competition for Arbitration Forums
Competition among different forums incentives them to work efficiently and resolve disputes fast. It also drives innovation and convenient services to increase their market share. Further, with opt-in arbitration and free market competition, many forums located in different countries could help to lower the fees structure.

A big shout out to EOS Newyork proposal “Free Market Competition for Dispute Resolution: {{Regarbitrator}}/{{Regforum}}”

Mediation by Technology

There will always be a huge influx of similar disputes on the network which could be used for analytics and further for preparing the technology-driven questionnaire to guide claimant and making the whole process convenient. The questionnaire could help to filter the disputes and aid mediation using Technology based on precedents of similar claims.

eBay follows a similar approach i.e in dispute a software collects the maximum data for evidence followed by mediation using the technology. Mediation is generally based on data derived from similar claims categorized into different categories like non-receipt, non-as-described etc. Different types help to structure pinpoint questions around claims, which assist with mediations and with filtration for human intervention.

How would it all be implemented on EOS?

Let’s try to understand the process of taking an example -

  1. Alice develops a dApplication and put a claimable bond (using Bond dApp) of value X naming an arbitration forum (which uses Technology as Forth Party).
  2. Bob (non-tech savvy) looking at dApplication bond and certificates, buys some service from the dApp. Later he finds out that the services(or product) wasn’t offered as promised and ends up in a dispute with Alice’s dApplication.
  3. Bob puts a claim against the Alice’s dApplications bond.
  4. The mentioned Arbitration forums start collecting the pieces of evidence and puts a questionnaire for Bob. Bob puts all the pieces of evidence on public forums.
  5. Looking at the evidences a software uses previous(similar) claims as precedent and put a suitable agreement/solution for Bob and Alice. If both parties agree to the solution, the claim gets resolved.
  6. Unresolved claim steps up for human intervention and forum assigns Human Arbitrator.
  7. If the award by arbitrator comes in favour of Bob, Arbitrator could enforce the ruling and transfer Token from the bond to Bob’s address without the need of Block Producer.

For More information on Tungsten, please watch —

Cryptolions are working on Bond dApplication as well, check out the interview to know more —

Please drop any questions/feedback/suggestion on our Telegram Channel: https://t.me/@eosbond

About EOS Metal

EOSMetal is a standby producer on EOS Network which uses bare metal servers /hybrid technology. Our only policy is one of transparency, we are self-funded and are dedicated towards maintaining top-class infrastructure for EOSIO ecosystem.

Learn more about EOSMetal