---
robots: noindex, nofollow
---
# XID Concept: Trust Metrics - Scoring and Propagation in a Progressive Trust System
## Expected Learning Outcomes
By the end of this document, you will:
- Understand the role of trust metrics in progressive and pseudonymous identity systems
- Learn how XIDs and Gordian Envelopes can encode, score, and propagate trust
- Explore practical strategies for trust computation in small, intentional online communities
- See how selective disclosure supports privacy-aware trust metrics
## Introduction
Traditional digital systems often rely on centralized authorities or binary trust decisions. These approaches lack nuance and fail to mirror how trust works in human relationships.
**Trust metrics** offer a computational model of reputation and reliability—often represented as scores derived from endorsements, attestations, and observed behavior. They support **progressive**, **context-sensitive**, and **transitive** trust building within pseudonymous systems like XIDs. When embedded in Gordian Envelopes, these metrics can be made verifiable, selectively disclosed, and privacy-preserving.
Trust metrics are especially critical in small, intentional communities where participants may not wish to reveal real-world identities but still need to build and signal trustworthiness.
> ⚠️ While trust metrics can help defend against sybil attacks and quantify behavior, they cannot resolve subjective disagreements or detect malice in socially complex scenarios. Community governance and social review remain essential complements.
## Principles of Trust Metrics
Key design principles for trust metrics in self-sovereign and privacy-preserving environments:
- **Transitivity with limits**: Trust may propagate indirectly, but must degrade over distance or be capped by a defined horizon.
- **Contextuality**: Trust is not global—it's bound to a domain (e.g., software contributions vs. moderation).
- **Subjectivity**: Agents may compute trust differently based on their preferences and starting points.
- **Transparency and verifiability**: All claims, endorsements, and scores can be cryptographically verified through Gordian Envelopes.
- **Sybil resistance**: Trust systems must resist manipulation by fake or duplicate identities.
## Integrating with XIDs
XIDs provide a stable, cryptographic identity layer to anchor trust metrics. Gordian Envelopes support layered and selective representation of trust assertions. This combination enables:
- **Direct trust**: Users can express ratings or endorsements for other XIDs.
- **Transitive trust**: Algorithms can compute derived scores based on trust graphs.
- **Contextual trust**: Each score can be scoped to a specific interaction type or domain.
Typical trust-building evidence includes:
- **Self-attestations**: Signed claims of skills or capabilities (e.g., “has contributed 20 patches”).
- **Peer endorsements**: Third-party signatures validating those claims.
- **Progressive trust envelopes**: Encapsulated lifecycle interactions based on [Progressive Trust](progressive-trust.md).
- **Interaction outcomes**: Logs of collaborative actions, audits, and dispute resolutions.
All these artifacts can be recorded and signed in envelopes, preserving both privacy and auditability.
## Example Trust Metric Models
### 1. Direct Endorsement Count
A straightforward model where trust increases with the number of distinct peer endorsements:
- Score = count of endorsements
- Optional: weight endorsements by the trust level or reputation of the endorser
### 2. Max-Flow Trust (Advogato-style)
Inspired by [Advogato’s trust metric](https://web.archive.org/web/20120516004623/http://www.advogato.org/trust-metric.html):
- Trust “flows” from seed nodes through a directed graph of endorsements
- Propagation is capped per edge to resist sybil attacks
- Algorithm computes a maximum flow from trusted sources to a candidate XID
### 3. Localized Trust (MoleTrust-style)
Based on [MoleTrust and TidalTrust by Jennifer Golbeck](https://drum.lib.umd.edu/handle/1903/3293):
- Personalized trust score computed from a specific source node
- Trust is propagated along endorsement paths within a defined depth (e.g., 2–3 hops)
- Trust decays with distance and endorsements further away hold less weight
> 📌 Horizon control is key: bounded propagation limits unintended influence from distant or loosely connected identities.
### 4. Reputation Aggregates
A composite model combining:
- Quantitative activity (e.g., number of code reviews)
- Qualitative assessments (e.g., peer feedback)
- Trust graph metrics (e.g., PageRank variants)
Useful for more complex governance and role-assignment systems.
> 📈 Trust graphs can be visualized using tools like Graphviz or Mermaid to help users understand trust propagation paths and identify central or disconnected nodes.
## CLI Examples
```sh
# List endorsements received by an XID
???
# Calculate trust score from a local perspective, limited to 2 hops
???
# Add a trustScore assertion to an envelope
envelope assertion add pred-obj string "trustScore" string "0.85" <ENVELOPE>
````
Trust assertions can also include metadata:
```sh
envelope assertion add pred-obj string "trustContext" string "Code Review" <ENVELOPE>
envelope assertion add pred-obj string "endorsedBy" string "ur:xid/hdcxalice" <ENVELOPE>
```
## Use Cases
* **Community onboarding**: Grant access to higher privileges based on trust scores or number of endorsements
* **Moderation systems**: Weight user reports or decisions by trust
* **Peer-to-peer delegation**: Enable re-endorsement or sponsorship chains
* **Conflict resolution**: Provide traceable history of interactions and trust justifications
## Privacy Considerations
Trust metric systems must be compatible with the privacy goals of XIDs:
* **Selective disclosure**: Users reveal only the endorsements or scores relevant to the interaction
* **Progressive trust reveal**: Additional trust data can be shared incrementally
* **Elision-friendly encoding**: Gordian Envelopes allow trust subgraphs or credentials to be omitted without breaking signature validity
## Future Directions
As progressive trust systems evolve, future trust metric systems might include:
* **Temporal decay**: Trust scores that decline over time without reaffirmation
* **Verifiable credentials**: Use of W3C-compatible VCs to encode endorsements
* **Zero-knowledge disclosure**: Prove trust level without revealing specific endorser identities
* **Inter-community bridges**: Federated trust models across multiple pseudonymous communities
* **Weighted dispute resolution**: Factor in trust scores when arbitrating disagreements or escalations
## Glossary
| Term | Definition |
| ------------------------ | ----------------------------------------------------------------------------------- |
| **Trust Score** | A numerical or categorical measure of reliability, often derived from endorsements |
| **Endorsement Graph** | A directed graph where edges represent trust or validation between XIDs |
| **Trust Horizon** | The maximum number of hops considered in trust propagation |
| **Decay Function** | A rule that reduces trust weight over time or distance |
## References
* [Progressive Trust Life Cycle](progressive-trust.md)
* [Fair Witness Assertions](fair-witness.md)
* [Pseudonymous Trust Building](pseudonymous-trust-building.md)
* **Advogato Trust Metric** — [archived page](https://web.archive.org/web/20120516004623/http://www.advogato.org/trust-metric.html) and [post-mortem discussion](https://web.archive.org/web/20170628190710/http://www.advogato.org/article/928.html)
* **MoleTrust and TidalTrust** — [Jennifer Golbeck dissertation (University of Maryland, 2005)](https://drum.lib.umd.edu/handle/1903/3293)
* **BrightID Sybil Resistance** — [https://www.brightid.org](https://www.brightid.org)
* Allen, Christopher. *[Musings on Progressive Trust](https://www.blockchaincommons.com/musings/musings-progressive-trust/)*
* Blockchain Commons. *[Progressive Trust Developer Guide](https://developer.blockchaincommons.com/progressive-trust/)*
Sign in with Wallet
Connect another wallet