Distributed consensus and benchmark evidence #
The Protocol Z.8 repository presents a ten-qubit star/GHZ circuit, reports a 98.85% logical-fidelity badge, and lists five claimed heartbeat runs averaging 92.42%. These are self-reported repository evidence. The records separate the classical majority rule and shot-derived bookkeeping from that claim and do not treat the repository prose as independent benchmark verification.
A finite majority decision for a measured bitstring.
A tie is represented by none; no tie-breaking convention is silently added.
Instances For
A strict majority of zeroes yields logical zero.
A strict majority of ones yields logical one.
A ten-node majority has the Protocol Z.8 threshold at five votes.
A ten-node majority has the Protocol Z.8 threshold at five votes.
A finite star council with one anchor and explicitly counted spokes.
- nodeCount : ℕ
- nodeCount_at_least_two : 2 ≤ self.nodeCount
- anchorCount : ℕ
- anchorCountLaw : self.anchorCount = 1
- spokeCount : ℕ
- spokeCountLaw : self.spokeCount + self.anchorCount = self.nodeCount
- globalEntanglingRounds : ℕ
- globalEntanglingRounds_positive : 0 < self.globalEntanglingRounds
- postProcessingMajority : Bool
Instances For
The star council exposes its one-anchor topology.
The star council exposes its node-count decomposition.
A shot-derived logical-fidelity observation.
logicalSuccesses counts accepted logical outcomes after the declared
post-processing rule. It is not a physical proof of fault tolerance.
- shots : ℕ
- shots_positive : 0 < self.shots
- logicalSuccesses : ℕ
- logicalSuccesses_le_shots : self.logicalSuccesses ≤ self.shots
- logicalFidelity : ℝ
- logicalFidelity_nonnegative : 0 ≤ self.logicalFidelity
- logicalFidelity_le_one : self.logicalFidelity ≤ 1
- logicalFidelityLaw : self.logicalFidelity = ↑self.logicalSuccesses / ↑self.shots
Instances For
The shot-derived logical fidelity is bounded by one.
The shot-derived logical fidelity is nonnegative.
A finite measurement report with a majority-decoded logical bit.
- council : StarConsensusCouncil
- measuredHammingWeight : ℕ
- hammingWeight_le_nodeCount : self.measuredHammingWeight ≤ self.council.nodeCount
- decodedBit : Option Bool
- decodedBitLaw : self.decodedBit = majorityDecision self.council.nodeCount self.measuredHammingWeight
- fidelity : ConsensusObservation
Instances For
The decoded bit is exactly the supplied majority decision.
One self-reported logical-fidelity run from the Protocol Z.8 heartbeat experiment.
- runId : ℕ
- logicalFidelity : ℝ
- logicalFidelity_nonnegative : 0 ≤ self.logicalFidelity
- logicalFidelity_le_one : self.logicalFidelity ≤ 1
- passed : Bool
Instances For
The five-run heartbeat table reported by the Protocol Z.8 repository.
The average is arithmetic bookkeeping over the five supplied percentages. It does not verify the hardware jobs or the reported decoder.
- backend : String
- date : String
- run1 : HeartbeatRun
- run2 : HeartbeatRun
- run3 : HeartbeatRun
- run4 : HeartbeatRun
- run5 : HeartbeatRun
- averageFidelity : ℝ
- averageFidelity_nonnegative : 0 ≤ self.averageFidelity
- averageFidelity_le_one : self.averageFidelity ≤ 1
- averageFidelityLaw : self.averageFidelity = (self.run1.logicalFidelity + self.run2.logicalFidelity + self.run3.logicalFidelity + self.run4.logicalFidelity + self.run5.logicalFidelity) / 5
- reportedAverageFidelity : ℝ
- reportedAverageFidelity_nonnegative : 0 ≤ self.reportedAverageFidelity
- reportedAverageFidelity_le_one : self.reportedAverageFidelity ≤ 1
- reportedAverageFidelityLaw : self.reportedAverageFidelity = 0.9242
Instances For
The heartbeat record exposes its five-run arithmetic average.
The heartbeat record exposes the rounded average printed by the source.
Evidence status for a reported consensus benchmark.
- modelOnly : ConsensusEvidenceStatus
- simulation : ConsensusEvidenceStatus
- reportedHardware : ConsensusEvidenceStatus
- independentlyReproduced : ConsensusEvidenceStatus
Instances For
Instances For
A reported Protocol Z.8 benchmark kept separate from shot-derived evidence.
The source claim is stored as provenance-bearing data. completeRawCounts and
reproduction make the missing evidence boundary visible instead of allowing
the claimed percentage to act as a theorem.
- backend : String
- heartbeat : HeartbeatExperiment
- jobIdentifier : String
- totalShots : ℕ
- totalShots_positive : 0 < self.totalShots
- majorityThreshold : ℕ
- majorityThreshold_positive : 0 < self.majorityThreshold
- consensusZeroCount : ℕ
- consensusOneCount : ℕ
- correctedCount : ℕ
- listedRawCountCoverage : ℕ
- listedRawCountCoverageLaw : self.listedRawCountCoverage = self.consensusZeroCount + self.consensusOneCount + self.correctedCount
- listedRawCountCoverage_le_total : self.listedRawCountCoverage ≤ self.totalShots
- reportedPhysicalFidelity : ℝ
- reportedPhysicalFidelity_nonnegative : 0 ≤ self.reportedPhysicalFidelity
- reportedPhysicalFidelity_le_one : self.reportedPhysicalFidelity ≤ 1
- reportedLogicalFidelity : ℝ
- reportedLogicalFidelityLaw : self.reportedLogicalFidelity = 0.9885
- reportedLogicalFidelity_nonnegative : 0 ≤ self.reportedLogicalFidelity
- reportedLogicalFidelity_le_one : self.reportedLogicalFidelity ≤ 1
- reportedCorrectionRate : ℝ
- reportedCorrectionRate_nonnegative : 0 ≤ self.reportedCorrectionRate
- reportedCorrectionRate_le_one : self.reportedCorrectionRate ≤ 1
- evidenceStatus : ConsensusEvidenceStatus
- completeRawCounts : Bool
- reproduction : Bool
- sourceRepository : String
Instances For
The Protocol Z.8 report exposes its claimed percentage.
A reported benchmark is not independently verified merely because it has a numeric fidelity field.
Instances For
A Protocol Z.8 report with incomplete raw counts cannot satisfy the local independent-verification predicate.