GenLayer Validators

These are the seven-day participation figures behind a validator's behavioural score. Think of a validator as a juror: these say how often it showed up and how often it agreed with the verdict.

Total votes

Rounds the validator took part in over the window. This is the sample size everything else is measured against.

Minority

Share of those votes that landed outside the consensus outcome. Lower is better; it drives the Consensus accuracy score.

Idle

Share of scheduled rounds the validator never showed up for. Distinct from Minority: there it voted and disagreed, here it did not vote at all.

As leader

Times the validator was picked to propose. It scores no points on its own; it decides how much confidence the Rotation figure is given.

Rotation

Share of those leader slots that had to be handed to someone else, because the proposal was late or did not carry.

Appeals won

Challenges the validator raised against an outcome and won. Informational only: it does not feed the score today.

Strikes

Disciplinary marks against the strike budget. At 3 the validator is tagged AT RISK; filling the budget means a ban.

Live

Whether the validator is reachable right now. Being offline multiplies the reliability score by 0.4 rather than shaving a few points off it.

Primed next

Whether it is registered for the coming epoch. Not being primed costs a flat 8 reliability points.

Below 60 votes the percentages are statistically meaningless, so every consensus-derived score is reported as n/a instead of a flattering number.

Where a validator's capital comes from, how much of it there is, and what it pays. All token amounts are already converted from base units.

Self stake

Capital the operator put up itself.

Delegated

Capital third parties delegated to it. The ratio between this and self stake is half of the Market trust score.

Delegators

How many distinct accounts delegated. A large delegated balance from very few accounts is a different risk than the same balance from many.

Weight

The validator’s voting weight in consensus. This is what the Nakamoto coefficient is computed over, not the stake figure.

APY

The nominal advertised yield.

Risk-adj APY

The nominal yield discounted by how well the validator actually behaves: APY × accuracy × security × reliability. Because it multiplies, one bad factor collapses the result.

The gap between APY and Risk-adj APY is the point of this table: the advertised yield is not the yield you collect. A jailed validator advertises a rate and delivers zero, and the risk-adjusted figure says so.

Five dimensions, each scored 0–100, then combined with fixed weights:

Accuracy · 30%

100 − minority%. Voting with consensus is the job, so it carries the most weight.

Reliability · 25%

100 − idle%, then × 0.4 if offline or banned, then − 8 if not primed.

Security · 20%

100 × (1 − strikes ÷ budget), or exactly 0 while banned.

Leadership · 15%

(100 − rotation%) × confidence, where confidence rises from 0.5 to 1 as leader slots approach 150.

Trust · 10%

Half stake size relative to the largest validator, half delegated-to-self ratio capped at 1:1 so capital alone cannot buy it.

A validator scoring 98.1, 96.6, 100, 94.4 and 25.4 on those five works out as:

Accuracy

98.1 × 0.3 = 29.43

Reliability

96.6 × 0.25 = 24.15

Security

100.0 × 0.2 = 20.00

Leadership

94.4 × 0.15 = 14.16

Trust

25.4 × 0.1 = 2.54

Composite

90.28

The confidence factor on Leadership is what stops a validator with three clean leader slots outranking one with two hundred: below 150 slots the score is damped toward 0.5. A validator that never led cannot have led badly, so it is treated as unproven rather than punished.

Cells turn amber past the thresholds the domain considers concerning: minority above 5%, idle above 5%, rotation above 12%.

The smallest number of validators whose combined voting weight exceeds half the network. Sort everyone by weight, add from the top, and count how many it takes to pass 50%.

It answers one question: how many parties would have to collude to control consensus? A coefficient of 3 means three operators are enough. Higher is more decentralised, and it is the single most honest number about a network's health: validator count says nothing if weight is concentrated in a handful of them.

It is computed over live voting weight, not stake and not a fixed supply, so it moves on its own as weight shifts between epochs.

No. This is an independent dashboard built and maintained by Stakely. It is not affiliated with, endorsed by, or published on behalf of GenLayer, and the ranking carries no official standing.

The score is indicative. Its weights are a judgement call about what makes a validator worth delegating to, not a protocol rule, and reasonable people would weight them differently, and a different weighting produces a different order. Every figure is derived from public explorer data, which can be stale, incomplete, or wrong.

Stakely runs a validator on this network and it appears in this ranking like any other, scored by the same code with no adjustment. Treat all of this as a starting point for your own research, never as financial advice.