Model deskSynthetic data
DevelopingFederal rescheduling: the broader proceeding is stayed. See the order and the next filing date.Official docket

Filed to this desk

Reporting

All reporting →

No published reporting has been filed to this desk yet.

Mojave Quant Terminal · model desk

Desk instrument

Full terminal →

Higher-order Fourier analysis · COA forensics

Nevada lab integrity leaderboard

‖f‖⁴U² = 𝔼x,h₁,h₂ Δh₁Δh₂f(x) = Σr |f̂(r)|⁴

Synthetic demonstration data. Six anonymized laboratories with planted behaviors (tier pinning, rounding, upward bias, suppressed microbial failures). A score is a screening signal that tells reporters where to pull COAs. It is never a finding of misconduct.

Labs flagged4/6review or escalate
Peer median THC23.32%median of lab medians
Pooled microbial fail4.8%all synthetic tests
Spectral spikes1above MP edge 1.54
#LaboratoryResultsIntegrityVerdictStrongest signal
01North Las Vegas8206▲ Anomalous — escalate to reportingIPR localization z > 8
02Reno59016▲ Anomalous — escalate to reportingMantissa entropy z > 8
03Pahrump52033◆ Review recommendedPotency inflation z 5.7
04Las Vegas70036◆ Review recommendedMicrobial suppression z 5.2
05Las Vegas76069● Consistent with natural variationTier step z 1.9
06Henderson64093● Consistent with natural variationMantissa entropy z 1.1

Synthetic Lab C · reported total THC, 820 results

  • Reported results
  • Smooth biological expectation
020406010%15%20%25%30%35%25% tier30% tier

Signal z-scores vs null

Gowers U²12.6
IPR localization23.2
Tier step12.9
Mantissa entropy5.8
Potency inflation3.3
Microbial suppression0.6

Rule marks z = 3, the review threshold.

Spectral anomaly radar

λ14.20
λ21.07
λ30.89
λ40.77
λ50.70
λ60.63

Rule marks the Marchenko–Pastur upper edge.

Leading-mode participation: Lab C 91%

Quant viewInspect the math & raw telemetry

Construction

Results are binned at 0.25 pt over 10–36% and smoothed with a Gaussian kernel. The residual f = (h − smooth h)/N is the shape left after removing biology.

Each statistic is compared with 40 null labs of the same size, drawn from the lab's own smoothed shape (continuous, unrounded, unpinned). The z-score asks how unusual this shape is for an honest lab of this size.

The composite is the root-sum-square of one-sided z-scores, each capped at 8. Integrity = 100 / (1 + (index/4)²). A lab that reports to one decimal by policy will trip the entropy test, so check reporting precision first.

Synthetic Lab C raw statistics

‖f‖U²
2.941e-3
IPR Σf⁴/(Σf²)²
0.2544
Tier step ratio
6.09
Mantissa entropy
0.9915
Median THC
25.41%
Microbial fail rate
4.4%
Anomaly index
15.38
LabGowers U²IPR localizationTier stepMantissa entropyPotency inflationMicrobial suppressionIndex
Synthetic Lab C12.623.212.95.83.30.615.4
Synthetic Lab D1.52.63.6155.00.5-1.09.3
Synthetic Lab F-0.10.1-1.0-0.25.7-1.15.7
Synthetic Lab E0.1-0.1-1.10.7-0.55.25.3
Synthetic Lab A0.01.71.90.7-0.7-2.52.7
Synthetic Lab B-0.2-0.00.01.1-0.7-1.51.1