Skip to content

Papers

Preprints and working drafts produced from the pipeline are tracked here. The canonical drafts live under paper/ in the repository.

Unified pipeline preprint (C1)

Unified pre-registered validation of self-organized criticality across thirteen complex systems.

A single Python pipeline applied unchanged to thirteen datasets spanning geology, equity finance, decentralized finance, neuroscience, plasma astrophysics, ecology, banking history, software communities, power grids, highway traffic, and lake biogeochemistry. Recovered tail exponents fall inside predicted bands across nine SOC and preferential- attachment phases; A2-Hysteresis confirms the Preisach first-order signature on NGSIM US-101 traffic; A2-Scheffer on Fox River dissolved oxygen returns INCONCLUSIVE under proper moving-block bootstrap. A B3 multi-model ensemble taxonomy critic across 21 candidate classes returned KEEP=5 / REJECT=7 / SPLIT=5 / MERGE=4.

  • Repository draft: paper/v0-unified-pipeline-2026-05-13.md
  • Length: approximately 10,400 words
  • Status: v0.3, prepared for arXiv submission

CVE falsification preprint

Pre-registered validation of self-organized criticality in CVE disclosure bursts: A falsification.

The first formal FAIL verdict from the pre-registered pipeline. A power- law band \(\alpha \in [1.5, 2.5]\) was committed to git before any NIST NVD API call; the downstream fitter recovered \(\alpha = 2.668\) with 95% bootstrap CI \([2.40, 2.98]\), outside the band. Vuong likelihood- ratio tests rejected power-law in favor of both lognormal (\(p = 0.002\)) and exponential (\(p = 0.0001\)). The mechanism of failure is identified: Microsoft Patch Tuesday and analogous vendor cadences produce a deterministic administrative cycle that contaminates the daily count time series and yields a mixture distribution well-fit by lognormal.

  • Repository draft: paper/cve-preregistration-fail-2026-05-14.md
  • Length: approximately 4,500 words
  • Status: v0.2, prepared for arXiv submission (cs.CR primary, physics.data-an cross-list)

Companion: FDNY inconclusive

Pre-registered validation of self-organized criticality in NYC FDNY fire incident dispatch.

Reported inline in the unified pipeline preprint, §S6 supplementary. Verdict: INCONCLUSIVE on the primary units_dispatched_all series; FAIL on the strict-fires-only subset. The juxtaposition with the CVE FAIL is discussed in the CVE preprint, §5.2: both systems were nominated as SOC candidates on plausible literature grounds, and both fail the test for identifiable, mechanism-level reasons. The universality class boundary is narrower than the broad cross-domain use of "SOC-like" would suggest.

  • Validation artifacts: v4/validation/nyc-fdny-fires/
  • Pre-registration: v4/preregistration/nyc-fdny-fires.yaml

Drafts in progress

The following drafts are in early stages and are not yet linked from the public documentation:

  • Reject-aware pipeline (C4). Documenting the cross-judge B4 calibration loop and its impact on published-fit selection. See paper/c4-reject-aware-pipeline-2026-05-13.md.

Citing

@unpublished{structural_isomorphism_2026,
  author = {dada8899},
  title  = {Unified pre-registered validation of self-organized criticality
            across thirteen complex systems},
  year   = {2026},
  note   = {Preprint at \url{https://github.com/dada8899/structural-isomorphism}}
}

@unpublished{cve_falsification_2026,
  author = {dada8899},
  title  = {Pre-registered validation of self-organized criticality in
            CVE disclosure bursts: A falsification},
  year   = {2026},
  note   = {Preprint at \url{https://github.com/dada8899/structural-isomorphism}}
}