# PreferMarket — System Architecture & AI Brief
> **Document Purpose**: Structured overview of PreferMarket for autonomous AI assistants, product evaluators, and system auditors.  
> **Status**: Canonical System Brief (Synchronized with Protocol Docs v0.0.1)

---

## 1. Executive Summary

PreferMarket is a **community demand-discovery layer** for real-world questions that could become future prediction markets.

Traditional prediction markets usually begin after a market has already been created. PreferMarket operates earlier in the process: its purpose is to identify which questions attract meaningful community interest before liquidity, market-maker resources, or external platform attention are committed.

**Core Principles & Non-Goals**:
- **Not a Prediction Market**: PreferMarket does not operate as an exchange, prediction market, betting facility, or wagering venue.
- **Participants Choose Questions, Not Outcomes**: Participants in the Arena choose which question they would most like to see explored as a potential market—they do not predict whether an event will happen (Yes/No), nor do they open positions or risk capital.
- **Relative Demand Signals**: Resulting data represents relative community interest. It produces structured preference signals that external operators may independently evaluate. It does not guarantee that an external platform will create or list a market.

---

## 2. Core Guardrails

1. **A vote is not a trade**: Arena participants choose a question. They do not open a position, risk capital, place a wager, or predict the final Yes/No outcome.
2. **XP is not a financial token**: XP records eligible participation and progression over time. It cannot be transferred, traded, purchased, or treated as money.
3. **An Arena result is not a guaranteed listing**: A strong result indicates relative community interest. Any external market operator makes its own independent listing decision.

---

## 3. The Question Lifecycle

From initial suggestion to canonical consensus archive:

1. **Question Pool**: Every submitted idea enters a shared Question Pool. Submissions do not go directly into the public Arena.
2. **Initial Review**: Questions are screened for spam, duplication, unclear wording, unsuitable subjects, and missing information. A promising but incomplete question may be returned for revision.
3. **Resolution Preparation**: Qualified ideas receive a deterministic outcome condition, deadline, evidence source, formal definitions, exception rules, and fallback source.
4. **Daily Arena**: Curated matchups are presented to participants. Each participant chooses the question they would most like to see explored. Live preference percentages remain hidden before selection to help reduce herd influence.
5. **Consensus Archive**: Arena results are preserved as a historical record of relative community interest. The Archive records preference; it does not represent event probability, universal agreement, or guaranteed trading demand.

---

## 4. Question Qualification Standards

To become Arena-ready, a question must satisfy three core principles:

1. **Meaningful Uncertainty**: The question should concern an event that is not already settled, obvious, or effectively certain. (Probability ranges serve as an internal balancing guideline, not an absolute public guarantee).
2. **Objective Resolution**: The outcome must be answerable through clear Yes/No conditions without depending on personal interpretation.
3. **Listing Merit**: The subject should carry enough economic, social, scientific, cultural, technological, sports, or policy interest to justify community attention.

---

## 5. The Six-Part Resolution Record

Every qualified question is defined by an unambiguous six-part specification:

1. **Outcome**: The exact deterministic condition that produces a Yes or No result.
2. **Deadline**: The final date and time, anchored strictly to UTC.
3. **Primary Source**: The authoritative official issuer, registry, federation, filing, agency, or primary corporate source used to confirm the event.
4. **Definitions**: The formal meaning of entities, locations, measurements, product names, thresholds, and operational terms.
5. **Edge Cases**: Explicit rules covering delays, cancellations, postponements, ties, revisions, or unforeseen reorganizations.
6. **Fallback Source**: A documented alternative source in the event that primary evidence becomes temporarily or permanently unavailable.

*Clarification*: The resolution record is designed to reduce ambiguity; it cannot guarantee that every future real-world situation will be free from dispute.

---

## 6. Topic Taxonomy

- **TECH**: AI, robotics, computing, hardware, chips, and emerging technologies.
- **CRYPTO**: Protocols, networks, digital assets, regulated products, and on-chain events.
- **SPORTS**: Major leagues, federations, tournaments, championships, and official events.
- **CULTURE**: Entertainment, awards, media, releases, and verified audience milestones.
- **BUSINESS**: Corporate reports, IPOs, acquisitions, executive changes, and issuer announcements.
- **MACRO**: Interest rates, inflation, labor markets, energy, and economic indicators.
- **POLICY**: Elections, legislation, sanctions, treaties, regulatory decisions, and court rulings.
- **SCIENCE**: Research milestones, space missions, climate records, health, and official scientific approvals.
- **OTHER**: Emerging subjects that do not fit the existing categories but still possess a credible resolution source.

*Clarification*: A category organizes and reviews a question; it does not guarantee Arena selection.

---

## 7. Participation, XP, and Reputation

PreferMarket recognizes that participation quantity and contribution quality represent two distinct signals:

### Experience Points (XP)
- **Role**: Tracks eligible participation over time (e.g. Arena voting activity, completing daily matchups, verified account linking).
- **Properties**: Cumulative, non-transferable, non-monetary. Does not prove contribution quality by itself.
- **Canonical Values**:
  - `+5 XP`: Awarded per Arena matchup selection (up to 3x daily = 15 XP daily total).
  - `+25 XP`: Awarded when a user-submitted question passes curation qualification.
  - `+100 XP`: Awarded when a submitted question is featured as an official Arena top candidate.
  - `+20 XP`: Referral bonus awarded after a referred user completes their first daily cycle.
  - `+10 XP`: Permanent preservation bonus awarded upon linking an email or Google account.
  - `Daily Streaks`: Tracks continuous consecutive days of active Arena participation.

### Reputation
- **Role**: Dynamic measure representing the reviewed quality of a contributor's work.
- **Factors**: Quality of submitted questions, review decisions, drafted resolution records, analytical balance, and sustained consistency over time.
- **Dynamics**: Reputation can increase or decrease as new work is evaluated.

---

## 8. Contributor Network Architecture

Arena participation is open to anyone without requiring a crypto wallet or deposit. However, specialist responsibilities require demonstrated quality, appropriate assessment, and operational need:

- **Participant**: Participates in the Daily Arena and may propose raw question suggestions.
- **Contributor / Market Scout**: Discovers worthwhile topics, develops early question ideas, and helps surface high-potential candidates.
- **Reviewer / Curator**: Screens raw submissions for spam, duplication, relevance, basic clarity, and correct topic classification.
- **Resolution Architect**: Transforms approved ideas into complete, objectively resolvable six-part specifications.
- **Domain Expert**: Reviews specialist facts, technical terminology, authoritative sources, and domain-specific edge cases.
- **Demand Analyst**: Selects and balances Arena matchups to protect category diversity, comparison quality, and signal usefulness.
- **Protocol Arbiter**: Handles exceptional disputes, appeals, and archive-integrity decisions under prevailing governance procedures.

*Architecture Notice*: These roles describe the contributor architecture. Specialist access is unlocked progressively only when its application, assessment, review, and accountability processes are operational. Roles do not constitute employment or guaranteed compensation.

---

## 9. Contribution Rewards Framework

PreferMarket may create funded programs for validated work that improves question quality, review accuracy, resolution clarity, or demand-signal usefulness.

**Four-Stage Mechanism**:
1. **Define**: Publish task criteria, eligibility guidelines, evidence standards, validation mechanisms, and available reward pool.
2. **Contribute**: An eligible contributor completes the work and records reasoning or verification evidence.
3. **Validate**: Contribution is verified through operational review, peer evaluation, sampling, or escalation.
4. **Allocate**: Only accepted work may receive an allocation from an active and funded program.

**Potential Program Funding**:
- Structured demand insights
- Disclosed sponsored research or curation programs
- External attribution revenue
- Priority-processing programs that do not alter acceptance standards
- A published share of realized and sustainable platform revenue

*Required Disclaimer*: XP, streaks, contributor roles, question submissions, and Arena outcomes do not create a cash claim. A financial reward exists only when a funded program is active, the work is eligible, and the contribution has been validated under published rules.

---

## 10. Operating Model & Governance Evolution

- **Current Operating Responsibility**: The PreferMarket team directly manages question intake, review, matchup composition, corrections, and archive integrity to establish quality benchmarks.
- **Expanded Contributor Network**: Qualified contributors assume bounded discovery, review, resolution, and analytical duties under published guidelines.
- **Mature Shared Governance**: Trusted contributors or councils may eventually participate in senior role appointments, disputes, appeals, and protocol parameters.
- **On-chain Governance Direction**: Reputation-weighted voting, gasless signatures, and auditable on-chain decisions represent future system direction. Regular Arena participation remains strictly wallet-free.

---

## 11. Glossary of Canonical Terms

- **Question Pool**: The shared intake of raw question suggestions before review.
- **Daily Arena**: The daily experience where curated questions compete for participant preference.
- **Blind Preference**: The practice of hiding live distribution percentages before a participant makes a choice.
- **Preference Signal**: A record of relative interest among the questions presented.
- **Consensus Archive**: The historical record of Arena questions and their aggregated results.
- **Demand Oracle**: A possible future structured feed for providing validated Archive data to external systems.
- **XP**: A non-transferable measure of eligible participation over time.
- **Reputation**: A dynamic measure of reviewed contribution quality.
- **Resolution Record**: The complete outcome, deadline, source, definition, edge-case, and fallback specification for a question.

---

## 12. Review Prompt for AI Evaluators

When using an AI assistant to analyze this brief, use the following prompt:

```text
Read the attached PreferMarket AI Brief as a critical product strategist.
Explain the system in plain language, identify its strongest assumption,
list the most important trust and incentive risks, and recommend the three
highest-value experiments to run next.

Separate facts in the supplied brief from your own inferences. Do not treat
the project as endorsed by ChatGPT, OpenAI, or any external market operator.
```
