Methodology
The reasoning behind our scoring framework.
Summary
Formula: TotalScore = 0.4 * Chain + 0.4 * Control + 0.2 * Fairness
Kill-Switch Cap: B5=0 caps total score at 1.0
Benchmark: Bitcoin is the reference for maximum decentralization
Why 40/40/20?
Chain Score (40%) measures technical decentralization - can the network survive attacks? This is fundamental: without a resilient network, nothing else matters.
Control Score (40%) measures governance decentralization - who actually controls the protocol? A technically decentralized network means nothing if one entity can change the rules, halt the chain, or steal funds through upgrades.
Fairness Score (20%) measures launch and distribution fairness. This matters less for ongoing decentralization but reflects the project's values and long-term incentive alignment. A 70% VC premine creates different dynamics than a fair launch.
Chain and Control are equally weighted because both are necessary conditions for decentralization. Fairness is weighted lower because it's a one-time historical fact that affects but doesn't determine ongoing decentralization.
Why Cap at 1.0 for Kill-Switches?
If a single entity can halt, freeze, or censor the chain (B5=0), the project is fundamentally not decentralized - regardless of how well it scores on other criteria.
A project with excellent validator distribution, diverse clients, and fair launch is still centralized if one admin key can pause everything. The kill-switch is a binary property: either you have one or you don't.
The 1.0 cap is harsh by design. It sends a clear message: "This project may have many good properties, but it cannot be considered decentralized."
We display the uncapped score for transparency, so users can see what the project would score without the penalty.
Why These 12 Criteria?
We selected criteria that are measurable, meaningful, and resistant to gaming. Each criterion captures a distinct aspect of decentralization that cannot be easily faked.
Chain Score (A1-A4)
- A1: Nakamoto Coefficient - Number of independent entities that would need to collude to compromise the system.
- A2: Validator/Miner Concentration - Share of top 5 validators/miners in stake/hashrate.
- A3: Client Diversity - Number of independent full-node implementations with relevant market share.
- A4: Node Geography & Hosting - Geographic distribution of nodes and cloud hosting concentration.
Control Score (B1-B6)
- B1: Corporate/Foundation Capture - Is there a dominant company/foundation controlling roadmap, marketing, and hiring? Can the project survive without them?.
- B2: Repo/Protocol Ownership - Distribution of merge rights in core repositories (clients, specs).
- B3: Brand & Frontend Control - Who owns brand, domains, main frontends, official wallets/apps? Decentralized ownership is better.
- B4: Treasury & Upgrade Keys - Composition of treasury/upgrade multisigs and admin keys.
- B5: Admin Halt Capability - Can a single entity or small group unilaterally halt, freeze, or censor the chain? This is a critical centralization risk.
- B6: Protocol Immutability - Has the protocol made fundamental rule changes (consensus mechanism, monetary policy, contentious forks)? Immutable rules are a core property of decentralization.
Fairness Score (C1-C2)
- C1: Launch Fairness / Premine - Team/VC/Foundation premine and launch model (fair launch vs.
- C2: Token Distribution & Governance Power - Share of insiders (team/VC/exchanges) in supply and governance votes.
Scoring Rubric
Each criterion is scored 0-10. Within each category, all criteria are equally weighted. Here's what the scores mean:
| Score | Meaning |
|---|---|
| 0 | Fully centralized / Single point of failure |
| 1-3 | Highly centralized / Few entities dominate |
| 4-6 | Moderately decentralized / Some concentration |
| 7-9 | Highly decentralized / Well distributed |
| 10 | Maximum decentralization / No single point of control |
Category Weights
- Chain Score: A1-A4 equally weighted (each 25% of Chain Score)
- Control Score: B1-B6 equally weighted (each ~16.7% of Control Score)
- Fairness Score: C1-C2 equally weighted (each 50% of Fairness Score)
N/A values (criteria that don't apply to a consensus type) are excluded from the average.
Example Calculation
Project with Chain=7.5, Control=6.0, Fairness=8.0:TotalScore = 0.4 × 7.5 + 0.4 × 6.0 + 0.2 × 8.0 = 3.0 + 2.4 + 1.6 = 7.0
Bitcoin as Benchmark
Bitcoin is our reference point for maximum decentralization. Not because it's perfect, but because it's the oldest, most battle-tested example of a truly decentralized network.
- No premine, no VC allocation, no foundation treasury
- No single entity can halt or censor the network
- No fundamental protocol changes since 2009
- Satoshi disappeared - the project survives without its creator
When evaluating any project, ask: "How does this compare to Bitcoin?"
What We Don't Measure
Decentralization is not the same as quality. We deliberately exclude:
- Price / Market Cap - Not a decentralization metric
- TVL - Measures adoption, not decentralization
- TPS / Performance - Often traded for centralization
- User Experience - Orthogonal to decentralization
- Team Reputation - Subjective and gameable
A highly centralized project can be fast, cheap, and user-friendly. A decentralized project can be slow, expensive, and clunky. We measure decentralization, nothing else.
Data Sources & Objectivity
We measure what's measurable and document transparently where judgment is required. Each criterion falls into one of two categories:
Control Score (B1-B6) - Expert Assessment
Requires human judgment. No API can determine if a foundation can halt a chain. Each assessment is documented with sources and reasoning.
- B1 - Corporate/Foundation capture - Who controls roadmap, hiring, marketing?
- B2 - Repo ownership - GitHub contributor analysis, merge rights distribution
- B3 - Brand & frontend control - Domain ownership, official apps, alternatives
- B4 - Treasury/upgrade keys - Multisig composition, signer independence
- B5 - Admin halt capability - Can anyone pause/freeze the chain?
- B6 - Protocol change history - Fundamental rule changes since launch
Fairness Score (C1-C2) - Mixed
Historical data that doesn't change. Verifiable from tokenomics sources.
- C1 - Premine % - Historical, verifiable (Messari, docs, blockchain explorers)
- C2 - Governance power - Snapshot voting data, token distribution analysis
Chain Score (A1-A4) - API Integration Planned
Currently manual assessment. API automation in development.
- A1 - Nakamoto Coefficient - Target: nakaflow.io API
- A2 - Validator concentration - Target: chain-specific APIs
- A3 - Client diversity - Target: node explorer APIs
- A4 - Hosting distribution - Target: bitnodes, datacenter analysis
For expert assessments, each value includes:
- Source links with dates
- Reasoning for the specific score
- Known limitations or caveats
- Full documentation in the project source files
No API can tell you if a foundation can halt a chain or if a team controls the brand. These require judgment - but that judgment is documented, sourced, and open to challenge.
Limitations
This framework is an approximation. Decentralization is a spectrum, not a binary. Some caveats:
- Data is updated periodically - snapshots, not real-time
- Some criteria require manual assessment (brand ownership, governance dynamics)
- New attack vectors may emerge that our criteria don't capture
- Scores reflect current state, not trajectory or potential
Use this as one input among many when evaluating projects. Not financial advice.