> For the complete documentation index, see [llms.txt](https://whitepaper.after-bell.uk/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://whitepaper.after-bell.uk/trust-model/administrative-resolution.md).

# Administrative Resolution

Explains the trust implications of administrative resolution decisions.

### An explicit exception path

Administrative resolution exists because AFTERBELL does not assume every market can safely resolve through the normal automated path. The normal path expects the Chainlink Stock Token Feed and the Robinhood REST API cross-check to be sufficiently close. When they differ by more than 5%, the protocol pauses the market rather than blindly resolving it.

The resulting path is an exception mechanism:

```
Oracle disagreement above 5%
              ↓
Circuit breaker
              ↓
Market pause
              ↓
Administrator uses off-chain evidence
              ↓
Manual resolution
              ↓
Winning outcome can be redeemed
```

This is not the normal resolution mechanism. It is the documented path for material disagreement between the primary price source and the secondary reference. It keeps the disagreement visible and introduces human judgment only when the automated comparison does not satisfy the stated condition.

### Trigger condition

`OracleResolver` reads Chainlink, cross-checks Robinhood, and applies the circuit breaker. The trigger is a difference greater than 5% between the two sources.

```
Chainlink price
       vs
Robinhood API price
       ↓
Difference >5%
       ↓
Circuit breaker → market pause → manual resolution
```

The source example is:

| Source    | Value |
| --------- | ----: |
| Chainlink |  $152 |
| Robinhood |  $140 |

The difference is approximately 7.9%. It exceeds the 5% threshold, so the market pauses. The normal automated resolution path does not select either value.

The threshold is a condition for pausing, not evidence that one source is correct. The source does not define a technical calculation method beyond the stated comparison rule. It also does not define other automatic triggers for administrative resolution.

### Why the market pauses

The pause exists because the protocol has detected a material disagreement between two reference values. Continuing automatically would treat that conflict as if it did not exist. The pause moves the market into an explicit exception state instead.

This choice has a clear trade-off. It can delay completion of the affected market, but avoids representing a disputed input as an unqualified automated result. The architecture favors stopping the normal path over blindly choosing the Chainlink value or the Robinhood API value.

The pause does not guarantee that the eventual result will be correct. It prevents the protocol from claiming automated certainty when the documented verification control has failed. The subsequent outcome depends on a manual review of off-chain evidence.

{% hint style="warning" %}
A paused market is not automatically resolved. The pause marks an unresolved disagreement and introduces an administrative trust boundary.
{% endhint %}

### Manual resolution

The source specifies that an administrator resolves the market manually using off-chain evidence. At a conceptual level, the process is:

1. A material oracle disagreement is detected.
2. The circuit breaker pauses the market.
3. An administrator reviews relevant off-chain evidence.
4. The administrator determines the resolution.
5. The market proceeds through the resolution path.
6. The winning outcome becomes redeemable.

The last step connects the exception path back to the on-chain outcome-token layer. `ConditionalTokens` represents Yes and No ERC-1155 positions and supports redemption. Once the resolution result is applied through the protocol path, the winning outcome retains redemption value and the losing outcome does not.

Manual review does not turn evidence into an autonomous on-chain fact. It is a human decision point that precedes the on-chain resolution and redemption state. The chain can enforce the resulting state transition. It cannot independently establish that the reviewed off-chain evidence was correct.

The source does not identify what evidence the administrator uses. It therefore does not support a claim about evidence sources, priorities, or review criteria.

### Human trust boundary

Manual resolution introduces a human trust assumption. In an exception case, users must rely on the administrator’s judgment to resolve the market from off-chain evidence.

This is a technical description of the system boundary, not a criticism of the mechanism. The design uses two data sources and a circuit breaker to reduce the chance of resolving a material discrepancy automatically. When that control detects a disagreement, the architecture explicitly retains a human path.

AFTERBELL is therefore not fully trustless at the resolution layer. The source states this directly: **“This is not trustless.”** The full model is “trust, but verify, twice”: primary Chainlink data, a Robinhood API cross-check, a pause above the threshold, and manual handling of unresolved disagreement.

### Normal and exception paths

| Path      | Trigger                                | Process                     | Trust dependency              |
| --------- | -------------------------------------- | --------------------------- | ----------------------------- |
| Normal    | Oracle disagreement is 5% or less      | Normal resolution           | Oracle data                   |
| Exception | Oracle disagreement is greater than 5% | Pause and manual resolution | Oracle data and administrator |

Manual resolution is not expected for every market. The normal path proceeds when the two values meet the 5% comparison condition. The exception path exists specifically because that condition may fail.

```
Normal
Chainlink + Robinhood API → disagreement ≤5% → resolution

Exception
Chainlink + Robinhood API → disagreement >5% → pause → administrator → resolution
```

Both paths still rely on external information. The exception path adds a human judgment dependency because the automated comparison did not permit normal resolution.

### End-to-end example

Consider “Will NVDA close above $150 on Friday?” The market has Yes and No outcome positions. Participants may have entered those positions through the trading and on-chain settlement path before the resolution condition occurs.

At resolution, Chainlink reports `$152` and the Robinhood API reports `$140`. Their difference is approximately `7.9%`, exceeding the 5% circuit-breaker threshold.

The normal automated path is interrupted. The market pauses. An administrator reviews off-chain evidence and manually resolves the market. After the resolution proceeds through the protocol path, the winning outcome can be redeemed through the outcome-token mechanics.

This example does not identify what evidence is reviewed, how long the review takes, or why the values differ. Those details are not defined by the current source. It illustrates only the documented transition from detected disagreement to pause, manual resolution, and eventual redemption.

### What is not yet specified

The current whitepaper does not specify the administrator’s identity or administrator count. It does not define a multisig architecture, governance architecture, voting rules, evidence hierarchy, appeal process, challenge period, or resolution SLA or deadline.

These omissions are current documentation boundaries. They should not be filled with assumed industry practices. A reader evaluating the exception path should recognize that the administrator and evidence-review mechanism are presently defined only at the level of manual resolution using off-chain evidence.

The source also does not define an automated recovery process that would replace the administrator once the circuit breaker has triggered. The documented response is manual, not an additional automatic oracle mechanism.

### Trust disclosure and current status

Administrative resolution is a deliberate exception mechanism. It improves the system’s ability to handle oracle disagreement, but it introduces human judgment. That judgment is part of AFTERBELL’s current trust model.

AFTERBELL is still under development. Smart contracts are not yet audited, and an audit is planned before mainnet. The design is documented in whitepaper **v0.1, September 2026** and is not production-proven infrastructure.

Revenue projections are not proven, and organic liquidity is not guaranteed. AFTERBELL does not currently have a token. It is a product, not a token launch.

{% hint style="info" %}
The administrative path does not replace the normal oracle path. It handles the defined case where the normal comparison detects disagreement above 5%.
{% endhint %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://whitepaper.after-bell.uk/trust-model/administrative-resolution.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.
