How It Works

How Trunod Works

A trust scoring system powered by probability theory, community validation, and real-time analysis — not a black-box algorithm.

1. Trust Analysis

Uses Bayesian probability to calculate trust scores. Evidence is weighted: vote ratio, account age, network signals, and Sybil detection.

2. Community Validation

Claims are validated by expert juries selected from the community. Transparent voting and consensus mechanisms ensure accuracy.

The Trust Score System

📊 Probability, not opinion

Trunod uses Bayesian probability theory to calculate realistic trust scores. This mathematical approach accounts for uncertainty and updates beliefs as new evidence emerges.

How it works step by step:

  1. 01
    Prior Belief: Every new user starts at a mathematically neutral 50% baseline — no assumptions made.
  2. 02
    Collect Evidence: Analyze votes, account age, network connections, and platform behavior.
  3. 03
    Weight Evidence: Strong signals (vote ratio) carry more weight than weaker ones (account age).
  4. 04
    Update Belief: Bayesian formula combines the prior with new evidence to produce the final score.
  5. 05
    Detect Threats: Sybil detection identifies fake accounts and bot networks before they influence scores.

Score Calculation Pipeline

Raw Signals
→
Evidence Weighting
→
Bayesian Update
→
Trust Score
Evidence Signals

🔍 What gets measured

Each signal contributes to the final trust score. The more positive signals, the higher the trust rating.

↑

Vote Ratio

Upvotes vs downvotes. High upvote ratio suggests trustworthy contributions.

Strongest signal
🛡

Account Age

Older accounts tend to be more trustworthy. New accounts start neutral.

Moderate signal
🌐

Network Signal

Connections to other trusted users increase credibility. Trust by association.

Strong signal
⚠

Sybil Detection

Pattern-based detection of fake accounts, bots, and coordinated networks.

Veto signal

Signal Strength Reference

Vote Ratio100%
Network Signal80%
Account Age50%
Sybil FlagVeto
Community Jury System

⚖️ Disputes resolved by experts

When claims are disputed, Trunod assembles juries of domain experts to validate or invalidate them. No single authority controls the outcome.

Jury Selection Process:

  1. 01
    Claim Disputed: A user or system flags a claim as potentially inaccurate.
  2. 02
    Topic Match: Jurors with verified expertise in the claim's domain (Tech, Finance, Science…) are identified.
  3. 03
    Random Selection: Jury members are randomly selected from the matched pool to prevent bias.
  4. 04
    Blind Voting: Jurors vote without seeing other votes — prevents groupthink and mob mentality.
  5. 05
    Consensus & Reward: Majority verdict wins. Accurate jurors earn Reputation Points; bad-faith disputes incur a penalty.

Dispute Resolution Flow

Claim Flagged
→
Expert Routing
→
Blind Vote
→
Consensus
→
Score Update

💡 Key safeguards:

  • ✓ No single authority controls truth
  • ✓ Expert opinions weighted by verified domain credentials
  • ✓ Blind voting prevents mob mentality
  • ✓ Transparent, auditable decision history
  • ✓ Failed Dispute Penalty deters weaponized reporting
  • ✓ Domain-specific credibility — a software engineer's score doesn't carry into medicine
Score Reference

Trust Score Ranges

0–20

Very Low

High risk

20–40

Low

Caution advised

40–60

Neutral

Insufficient data

60–80

Good

Generally trustworthy

80–100

Excellent

Highly trustworthy

Ready to see trust in action?

Install the Trunod extension and start viewing real-time trust scores on your favorite platforms.

Get Started FreeAdd to Chrome