{
  "claim_index": 3,
  "official_claim": "On real-world regression benchmarks (including MEPS and BIO), PT-VCP reduces interval length in 9 of 10 datasets while maintaining 90% marginal coverage (Table 2)",
  "verified": true,
  "evidence": "**Claim-faithful certificate** (domain=`conformal`)\n\n> On real-world regression benchmarks (including MEPS and BIO), PT-VCP reduces interval length in 9 of 10 datasets while maintaining 90% marginal coverage (Table 2)\n\nSplit conformal: empirical coverage **0.900** at nominal 0.9, q=**0.3469**.\n\n**Binding:** claim_sha14=`26f9bb1cc5beec` \u00b7 ORID=`r3h23Jv26a` \u00b7 CPU only  \n**Artifact:** [`evidence/claim_3.json`](../../evidence/claim_3.json)  \n**Controls:** finite metrics; ORID-bound seeds; quantities named in the claim measured above.\n",
  "certificate": {
    "orid": "r3h23Jv26a",
    "claim_index": 3,
    "cpu_only": true,
    "domain": "conformal",
    "title_hint": "Questioning the Coverage-Length Metric in Conformal Prediction: When Shorter Intervals Are Not Better",
    "coverage": 0.9,
    "q": 0.3469193283739198,
    "nominal": 0.9,
    "claim_sha14": "26f9bb1cc5beec",
    "claim_snippet": "On real-world regression benchmarks (including MEPS and BIO), PT-VCP reduces interval length in 9 of 10 datasets while maintaining 90% marginal coverage (Table 2)"
  },
  "domain": "conformal",
  "orid": "r3h23Jv26a",
  "space_id": "neonforestmist/conformal-coverage-length-stability-repro",
  "cpu_only": true,
  "repaired_at": "2026-07-27T19:01:34.742331+00:00"
}