Our Commitment to Transparency
At Scamtrustor, we believe every Trust Score should be explainable. Unlike black-box AI systems, our scoring engine provides detailed reasons for every score — reasons you can read, question, and understand.
The Six Scoring Pillars
Our Trust Score (0-100) is calculated across six weighted categories:
- Review Quality (35%): Bayesian-weighted, time-decayed ratings from verified community members
- Authenticity (20%): Anti-fraud signals, review pattern analysis, behavioral signals
- Identity (15%): Domain age, WHOIS data, registration patterns, business verification
- Security (15%): SSL certificates, HTTPS, security headers, DNS health
- Transparency (10%): Claimed profile, owner responsiveness, public information quality
- Performance (5%): Uptime, response time, reliability signals
Why Separate Risk Score?
The Trust Score reflects community trust and positive signals. The Risk Score (also 0-100, higher = more dangerous) captures negative fraud signals independently. A site might have a moderate Trust Score but a very high Risk Score — meaning some users trust it, but our systems detect serious danger signals.
Fraud Probability
Our fraud probability (0.0 to 1.0) is calculated from a weighted combination of behavioral fraud signals: review manipulation, infrastructure shared with known scams, sudden reputation changes, and more. A probability above 0.7 triggers automatic review flags and admin notification.