Tyler retweetledi

DECENTRALIZATION OF L1 POS BLOCKCHAINS
A COMPREHENSIVE ANALYSIS 2025
ETH | SOL | TRX | ADA | DOT | AVAX | SUI
When the media and influencers, when they talk about decentralization, talk only about the "Nakamoto Coefficient" - yes, it is important. And no, it is not the only parameter of decentralization. That's why we're starting right now.
I must say right away that I will not publish all the calculations here, it would take too much of your time. This is my personal opinion - I do not claim to be 100% true. DYOR.
Let's take as a basis that the total score in the table is 100% = this is the maximum possible decentralization that everyone is striving for. But so far no one has achieved it and probably won't, and I doubt it's possible. That's why even the best decentralized blockchain doesn't have 80%.
This analysis evaluates the decentralization of Layer 1 PoS (Proof of Stake) blockchains through 8 layers, based on the methodology of EDI (Edinburgh Decentralisation Index) - I took it as a basis, but it's not the same (!). Each layer focuses on a specific aspect of the network, using metrics like HHI, Gini, Shannon Entropy, and Nakamoto Coefficient. Importance: In PoS, decentralization reduces risks of attacks, censorship, and failures, increasing resilience. The layer weight reflects its contribution to the overall score (0-100). Below is a brief overview of each layer, with parameters and their importance in evaluating PoS blockchains.
Hardware (5%)
This layer evaluates the concentration of hardware infrastructure (validator hosting) to identify risks of provider failures.
1. HHI by cloud providers: measures concentration across cloud providers (high HHI = risk from dominance, e.g., AWS). Importance: In PoS, a provider failure can stop validators, reducing liveness.
2. Dependency ratio on top providers: share of top-3 providers. Importance: Shows dependence on a few, vulnerable to attacks or downtime.
Software (7%)
Focus on diversity of software clients (node implementations) to avoid bugs in a dominant client
1. Concentr. ratio for top 2 clients: Share of top-2 clients. Importance: In PoS, dominance (e.g., Geth in ETH) risks a fork from an error.
2. Shannon entropy by client shares: diversity of client shares. Importance: high entropy = more alternatives, increases resilience to vulnerabilities.
Network (8%)
Evaluates network decentralization (connectivity) to make the network resilient to DDoS or partitions
1. HHI by ASNs: concentration across autonomous systems (ASNs). Importance: In PoS, high HHI in ASNs (e.g., AWS ASNs) makes the network vulnerable to ISP failures
2. Gini by node peers: Inequality in the number of peers per node. Importance: even peer distribution ensures good gossip, reducing latency and censorship
3. Shannon entropy by node peers: Diversity of peers. Importance: High entropy = network not dependent on a few hubs, improves liveness
Consensus (34%)
Key layer (!) - concentration in block production, determines PoS security.
1. Nakamoto coefficient: minimum number of entities to control >1/3 stake. Importance: Shows attack risk (low NC = centralization).
2. Gini coefficient: stake inequality. Importance: High Gini = concentration, vulnerability to whales.
3. Effect.number validators/pools: effective number of entities (exp(Entropy)). Importance: Accounts for real diversity, balancing raw count.
4. Cost of 33% attack: cost of 1/3 stake attack. Importance: Economic protection in PoS, high cost = low attack feasibility.
5. Shannon entropy: stake diversity. Importance: High entropy = evenness, reduces risks.
6. Herfindahl-hirschman Index: stake concentration. Importance: Penalizes large entities, standard for monopoly risks in PoS.
7. "1-concentration ratio": share of top-1. Importance: shows dependence on one entity.
Tokenomics (24%)
Token distribution and economics, influencing incentives
1. Gini by holders: Inequality among holders. Importance: High Gini = concentration in whales, manipulation risk.
2. HHI by token distribution: token concentration. Importance: High HHI = economic power in few hands, low decentralization.
3. Voter turnout rate: % participation in staking/governance. Importance: High turnout = active community, reduces speculation in PoS.
API Clients (2%)
Concentration of API providers for network access
1. HHI by API providers: API concentration. Importance: High HHI (e.g., Infura dominance) = risk of outage for dApps
2. Dependency ratio: Share of top-3 API. Importance: Dependence on a few = single point of failure for users.
Governance (13%)
Diversity of governance for democracy
1. Shannon entropy by proposals: diversity of proposals by authors. Importance: High entropy = proposals from many, reduces elite control.
2. Gini by votes: Inequality of votes. Importance: High gini = votes concentrated, low participation diversity.
Geography (7%)
Geographical distribution for censorship resilience.
1. HHI by countries: concentration by countries. Importance: High HHI = risk from regional regulations/outages
2. Censorship resistance Index: share in censored regions. Importance: Low index = high censorship risk in PoS.
Why Nakamoto coefficient differs from nakaflow website? The Nakamoto coefficient on "nakaflow io" is higher because they divide pools into multiple node operators (e.g., ETH: Lido - ~500 operators, not 1), which increases the number of entities and NC. This is a more optimistic approach, but it can overstate decentralization if operators are connected (e.g., through cloud). I prefer to consider entities
Conclusion:
This is not "just random numbers" - this is a structured analysis (NC, Gini, HHI, etc.), which help compare chains in terms of security, resilience, and risks (e.g., 33% attack, stake concentration). Usefulness: shows strengths/weaknesses (e.g., Cardano high in distribution, ETH low in NC), but due to approximations (old data for hardware/network, N/A for governance) this is not "final", but indicative. In my opinion, $ADA is the most mature blockchain in terms of decentralization, but there is room for growth. For example, develop alternative Software to reduce concentration. I also see some problems with SOL and ETH, especially in the Governance layer. TRX looks the worst of all. DOT AVAX SUI shows good metrics.
FYI: none of the blockchains paid me for this research, so it's as independent as possible.
What kind of research should I do next? maybe for a complete picture of the blockchain trilemma, it is necessary to consider the scaling and security of PoS blockchains in separate studies.

English




























