The Nakamoto Coefficient (or Minimum Attack Vector, MAV) represents the minimum number of independent entities (or pools) required to control over 50% of the active stake. In Cardano's Ouroboros consensus, controlling >50% is required for a majority consensus attack (~20–24 entities when clustering multi-pool operators, or ~28–35 on individual pool level). Detailed methodology can be found on our Nakamoto Coefficient Live Tracker.

Cross-chain comparison platforms often apply a 33.3% PoS Liveness threshold (the BFT stalling point) across all networks, whereas Cardano's Ouroboros consensus requires a 50% majority threshold for consensus reorganisation. To control 33% of Cardano stake, only ~15–16 entities are needed, while a true >50% majority requires ~20–24 entities.

Cardano Observer aggregates stake pools into unified entities to prevent Multi-Pool Operators (MPOs) from artificially inflating decentralization metrics. Our clustering algorithm groups pools by:

1. Explicit Pool Group Metadata (e.g., Binance, Coinbase, Wave, 1PCT).
2. Shared Owner Stake Credentials (matching primary operator reward/signing keys).
3. Identical Reward Account Addresses (where epoch margins are pooled).

This clustering reveals that controlling >50% of active stake requires colluding with ~20–24 independent entities.

To guarantee data integrity, Cardano Observer enforces a strict validation rule: decentralization metrics (Nakamoto Coefficient, Gini Coefficient, and Entity MAV) are only computed if the network query returns at least 1,000 active pool records. If an upstream API pagination fails or returns a truncated pool list, the partial sample is rejected and the last fully validated cached snapshot is served from Cloudflare KV.

Cardano's smart contract infrastructure has evolved across three major eras:

• Plutus V1 (Alonzo, 2021): Introduced foundational on-chain smart contract validation and the eUTXO scripting model.
• Plutus V2 (Vasil, 2022): Introduced CIP-31 (Reference Inputs to read datums without spending UTxOs), CIP-32 (Inline Datums stored directly on-chain), and CIP-33 (Reference Scripts to reuse script bytecode by hash), reducing transaction sizes and fees by up to 90%.
• Plutus V3 (Chang / Conway, 2024): Introduced native Bitwise primitives for high-performance cryptography, BLS12-381 elliptic curve pairing for zero-knowledge proofs (ZKPs), and integration with CIP-1694 governance voting contexts.

Track live era distribution on the Plutus Telemetry page.

Execution Units (ExUnits) measure the computational effort of evaluating a smart contract. They consist of CPU Steps (execution clock ticks) and Memory Units (RAM consumption). Cardano enforces strict bounds:

• Per-Block Limit: 40,000,000,000 CPU steps (40B) and 62,000,000 Memory units (62M).
• Per-Transaction Limit: 10,000,000,000 CPU steps (10B) and 14,000,000 Memory units (14M).

These limits ensure that block validation times stay well below the 3-second block propagation threshold across global relay nodes.

Unlike account-based blockchains where transactions compete in a priority gas-fee bidding auction, Cardano node mempools operate on a deterministic First-In, First-Out (FIFO) queue. Slot leaders include valid transactions in the order they arrive. Because users cannot pay higher fees to jump ahead in the queue, predatory Miner-Extractable Value (MEV) tactics—such as front-running and sandwich attacks—are eliminated at the consensus layer. View real-time mempool telemetry on the Mempool Tracker.

Block propagation delay measures the time elapsed (in milliseconds) between when a slot leader mints a block and when that block is received, verified, and adopted by $\ge 95\%$ of global relay peers (P95 latency). In Cardano Observer, the empirical diffusion model incorporates base peer serialization (0.48s), payload weight (16ms/kB), script memory budget (22ms/MB), transaction count (3ms/tx), and geographic mesh jitter (0.04s–0.22s). Track live latency metrics on the Block Propagation Telemetry page.

• Active Stake: The snapshot of delegated ADA taken at the boundary of epoch $e-2$. This frozen figure determines the pool's VRF slot leadership schedule and block minting rights for the current epoch $e$.
• Live Stake: The real-time sum of ADA currently residing in all delegator wallets in the current epoch. Live stake reflects immediate delegation changes, deposits, or withdrawals and will become active two epochs later.

Pool pledge represents the operator's financial commitment to their pool. If an operator withdraws ADA from their owner reward keys such that Live Pledge falls below Declared Pledge, the pool enters an Unmet Pledge status. In this state, the pool receives zero delegation rewards for that epoch, penalizing both the operator and delegators.

The saturation limit is the maximum active stake a single pool can hold before reward payouts are capped. It is dynamically calculated as:
Slimit = Circulating Supply / k
With $k = 500$, the saturation limit is currently ~72–87M ADA. If a pool exceeds this limit, total epoch rewards do not increase, which dilutes the Return on Stake (ROS) for delegators and financially incentivizes them to redelegate to smaller pools.

At every 5-day epoch boundary, approximately $\rho = 0.3\%$ of the remaining unminted ADA in the protocol reserve is released into the reward pot alongside all transaction fees collected during that epoch. 20% of this combined pot ($\tau = 0.20$) is allocated directly to the on-chain Treasury, while the remaining 80% is distributed to pool operators and delegators, yielding an annual Return on Stake (ROS) of ~2.6%–3.2%.

Check our comprehensive Cardano Network Metrics Glossary for full definitions of eUTXO, Plutus budgets, slot leadership, MPO clustering, and stake distribution.