Network & decentralization metrics
Validator-level metrics describe how individual actors behave. Network-level and decentralization metrics describe the environment in which they operate: how stake is distributed, how concentrated control has become, and how the overall system evolves over time. FortisX combines both perspectives when assessing risk and designing allocation policies.
This section outlines the main categories of network-level metrics that FortisX maintains across validator-based networks such as Ethereum, Solana, Polkadot, Avalanche, and Cosmos, and how these metrics are used within the platform.
Objectives of network-level analytics
Network and decentralization metrics in FortisX are designed to answer questions such as:
How is stake distributed across validators, pools, and providers?
How concentrated is effective control in practice, and how is this changing over time?
How active is the network in terms of participation, governance, and protocol-level events?
Are there emerging patterns that may affect the risk of running or delegating stake within a given network?
These questions are addressed through metrics that capture participation, stake distribution, concentration, churn, and indicators related to governance and protocol changes.
Participation and staking participation
At the network level, FortisX tracks overall staking and participation characteristics, including:
Total staked amount – the aggregate quantity of stake that is actively participating in validation according to network rules.
Staking participation ratio – the ratio of staked stake to relevant supply measures (for example, a subset of circulating or eligible supply, depending on network definitions).
Effective participation – how much of the staked set is actively contributing to consensus versus being nominally staked but frequently unavailable.
These metrics help characterise how much of a network’s potential validation capacity is actually used, and how sensitive the network might be to concentrated failures or participation drops.
Stake distribution across validators, pools, and providers
The distribution of stake is a central input for decentralization and concentration analysis. FortisX maintains time-series metrics on:
Stake share by validator – each validator’s percentage of the network’s total active stake.
Stake share by pool – in networks or setups where pools are present, the share of stake aggregated at each pool.
Stake share by provider – aggregated stake across validators and pools operated by the same provider, across one or multiple networks where relevant.
These metrics are computed at regular intervals and stored historically, so that FortisX can observe how distribution patterns change, whether some entities gain disproportionate share, and how quickly those changes occur.
Concentration and decentralization indicators
Using stake distribution metrics, FortisX derives indicators that describe how concentrated or diffuse control in a network appears to be. Examples of such indicators include:
Share of stake in top N validators, pools, or providers – for several values of N, providing a simple view of how much stake is held by the largest actors.
Concentration indices – such as measures analogous to the Herfindahl–Hirschman Index (HHI) or other dispersion metrics computed over stake shares.
Tail distribution characteristics – such as how quickly stake share declines beyond the top tier of validators or pools.
These indicators do not attempt to define “good” or “bad” levels of decentralization in isolation. Instead, they provide a quantitative basis for comparing networks, monitoring changes within a network, and expressing policies that depend on concentration levels.
Churn and stake mobility
Static snapshots of stake distribution only show part of the picture. FortisX also tracks churn and stake mobility:
Validator and pool entry/exit rates – how often validators or pools enter or leave the active set.
Stake inflows and outflows at the network, pool, and provider levels over defined time windows.
Persistence of stake allocations – how stable stake positions have been, and whether there are frequent reallocations between a small number of destinations.
Churn metrics help identify environments where stake is highly mobile or where a few destinations periodically absorb disproportionate inflows. This has implications for both operational risk and for the stability of allocation policies.
Activity, load, and protocol events
In addition to stake distribution and concentration, FortisX observes indicators related to overall network activity and load, for example:
Transaction and block activity metrics, where relevant to the performance and reward environment for validators.
Governance and protocol events, such as parameter changes, upgrades, or votes that may affect staking conditions or validator operations.
Network-level incident signals, such as extended finality delays, elevated reorganisation rates (if applicable), or coordinated responses to known issues.
These indicators are used to provide context for validator and stake behaviour. For example, a sudden change in stake concentration may follow a protocol event, or an increase in minor penalties may correlate with a period of elevated network stress.
Cross-network comparability
Different networks use different units, epochs, and mechanisms. FortisX does not attempt to flatten these differences into a single synthetic score, but it does align metrics sufficiently to support structured comparisons:
core concepts such as stake share, concentration indices, and churn rates are defined in a network-agnostic way;
network-specific details (such as epoch length, penalty regimes, or governance mechanisms) are recorded and used to interpret metrics correctly;
derived indicators can be grouped into common dimensions (for example, concentration, churn, activity), while preserving per-network nuances in documentation and metadata.
This approach enables policies that impose similar structural constraints across networks (for example, limits on acceptable concentration levels) without assuming that all networks behave identically.
Data sources and quality considerations
Network-level and decentralization metrics are derived from:
on-chain state and events obtained from full nodes or indexers;
observed validator sets and active stake distributions at relevant intervals;
protocol metadata and governance data where accessible and reliable.
As with validator metrics, FortisX records the provenance of network metrics and applies consistency checks where multiple sources are available. When upstream data is incomplete or ambiguous, the platform may:
mark certain metrics as degraded or provisional;
delay the use of specific indicators in risk or policy evaluations until data quality improves;
record data quality status alongside metrics for downstream consumers.
This ensures that allocations and risk assessments can take into account both the measured properties of a network and the reliability of the measurements themselves.
Use within FortisX
Network and decentralization metrics flow into several parts of the platform:
Analytics and monitoring – providing views of how each supported network evolves in terms of stake distribution, concentration, churn, and activity.
Risk modeling – contributing to factors that reflect concentration risk, governance and protocol stability, and environment-level conditions that affect validators and pools.
Policy design and evaluation – enabling policies that:
restrict exposure to networks with certain concentration characteristics;
require minimum levels of decentralization or participation for eligibility;
adapt allocation decisions in response to sustained changes in network structure.
Because network-level metrics are recorded as time series and tied to specific model and policy versions, FortisX can reconstruct which network conditions and assumptions were in place when particular allocation decisions were made. This is important for both internal governance and external review.
The next sections describe how these metrics combine with validator-level data in the risk modeling layer, and how the resulting signals are used by the policy engine to shape allocations and rebalancing behaviour.
Last updated