{
  "abstract": "Trimming suspicious calibration points is a common response to contamination in conformal prediction. Its effect on clean-target coverage, however, is governed by the retained law induced by trimming, not by the contamination level alone. We analyse fixed-threshold trimming as conditioning rather than purification. It replaces the contaminated calibration law with a retained law, reducing clean-target coverage to a one-dimensional score-CDF transfer problem with an exact finite-sample identity. A componentwise bound on the transfer gap gives a population-level diagnostic. This separates a clean-side covariance cost from a retained-contamination cost, governed by the dirty-to-clean retention ratio. Trimming helps when the anomaly score separates retention probabilities while remaining score-neutral on the clean population. Otherwise, it cannot substantially reduce contamination through the retained mixture coefficient. We also give finite-sample certificate templates that provide numerical guarantees under independent audit.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": "2605.06204",
  "author_total": 1,
  "authors": [
    {
      "name": "Congye Wang"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/7xcwxxg3a4aj2.bib",
  "cited_by": 0,
  "doi": "10.48550/arxiv.2605.06204",
  "fulltext_url": null,
  "id": "7xcwxxg3a4aj2",
  "issns": [],
  "landing_url": "https://arxiv.org/abs/2605.06204",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "http://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": "CC0-1.0",
  "native_subjects": [
    [
      "arxiv",
      "cs.LG"
    ],
    [
      "arxiv",
      "stat.ML"
    ],
    [
      "openalex",
      "1702|Adversarial Robustness in Machine Learning"
    ]
  ],
  "openalex": "W7160608758",
  "pdf_url": "https://arxiv.org/pdf/2605.06204",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-12",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/7xcwxxg3a4aj2.ris",
  "sources": [
    {
      "first_seen": "2026-10-30T12:29:26Z",
      "native_id": "2605.06204",
      "source": "arxiv",
      "url": "https://arxiv.org/abs/2605.06204"
    }
  ],
  "subjects": [
    {
      "field": "mathematics",
      "origin": "arxiv:stat.ML",
      "subfield": "statistics"
    }
  ],
  "title": "When Does Trimming Help Conformal Prediction? A Retained-Law Diagnostic under Calibration Contamination",
  "title_text": "When Does Trimming Help Conformal Prediction? A Retained-Law Diagnostic under Calibration Contamination",
  "type": "preprint",
  "updated_at": "2026-10-30T12:29:26Z",
  "url": "https://research.draconitic.com/paper/7xcwxxg3a4aj2",
  "venue": "arXiv",
  "versions": []
}