Research
BREM is built on peer-reviewed research, not marketing claims. Our methodology, dataset, and scoring instrument are all publicly available.
IEEE ICBC 2026 Paper
AI-R2D2 Workshop — Submitted March 2026
Mutation Authority as a Predictive Risk Factor: Empirical Evidence from 83 Blockchain Deployments
Todd Price — Real Blockchain Solutions
We present the Blockchain Risk Evaluation Model (BREM), a 9-cell risk matrix derived from the Structure-Process-Persistence framework. Validated against 83 real-world blockchain deployments (71 enterprise, 12 DeFi), BREM achieves F1=0.929 using a domain ceiling rule. The key finding: Mutation Authority (sm) — measuring who can change protocol rules — is the single strongest predictor of project failure. No project with sm ≤ 2 has failed in the dataset.
Open Dataset
83 projects · CC BY 4.0
The complete BREM dataset includes cell-level scores, domain aggregates, outcome classifications, and evidence summaries for all 83 projects. Enterprise (71) and DeFi (12) projects spanning 2016-2026.
View on GitHub →Key Findings
Threshold Performance
Overall score ≥ 3.0 (peak-sensitive formula) achieves 97.5% recall with F1=0.876 across the 83-project dataset. The peak-sensitive formula prevents low law scores from masking concentrated architectural risk.
Domain Ceiling Rule
Adding a domain ceiling check (any domain > 3.0) improves sensitivity to 97.5% (F1=0.929), catching "technical zombies" masked by strong legal or economic profiles.
Asymmetric Cell Weighting
High cell scores are strictly more impactful than low scores. Consensus risk carries 8.2x asymmetry — being bad at consensus is far more dangerous than being good at it is helpful.
The DeFi Reversal
Enterprise failures cascade Architecture → Scalability → Economics. DeFi reverses this: Economics → Mutation Authority. Token-price collapse exposes governance centralisation.