Resolution Is Not Accessibility: Coupled Minimax Risk for Reduced Statistical Experiments

Yuzhao Zhu, Jiahui Shao, Dongqi Liu, Xiaoxiao Hu

arXivarxiv/2611.02914Nov 2, 2026math.STscore 0

A reduced experiment can have vanishing minimax risk yet lose arbitrarily much precision relative to full observation. We determine when growing reductions preserve target-specific minimax precision, their exact cost, and which omissions can remain without increasing that cost. For Gaussian sequence reductions, we give necessary and sufficient criteria for bounded relative risk and losslessness. With parameter-dependent information, sampling difficulty and unresolved target variation draw on the same parameter budget. We prove that the exact global asymptotic minimax risk is the maximum…

SourceJSON

Comments and reviews

No comments yet.

API

curl /api/v1/records/arxiv/2611.02914
curl -X POST /api/v1/records/arxiv/2611.02914/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2611.02914/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"body": "…", "review": true}'