{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:NRYTMQQM5RQZRHWZ5UBON7Z3OC","short_pith_number":"pith:NRYTMQQM","schema_version":"1.0","canonical_sha256":"6c7136420cec61989ed9ed02e6ff3b70be2dd491e438af8d0314d97819ab23df","source":{"kind":"arxiv","id":"2606.29771","version":1},"attestation_state":"computed","paper":{"title":"CLQT: A Closed-Loop, Cost-Aware, Strategy-Consistent Benchmark for Diagnostic Evaluation of LLM Portfolio-Management Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","q-fin.CP","q-fin.PM"],"primary_cat":"cs.AI","authors_text":"Bo Qu, Mingguang Chen","submitted_at":"2026-06-29T04:26:30Z","abstract_excerpt":"LLM agents are increasingly cast as autonomous portfolio managers, and benchmarks have moved from financial question-answering to sequential trading. Yet most still rank agents by returns over a fixed window -- a weak proxy, since a period's return is dominated by the market path and apparent alpha can dissolve once look-ahead leakage is controlled. Such a ranking certifies neither sound reasoning, nor a consistent strategy, nor a durable edge. We introduce CLQT, which reframes closed-loop trading evaluation as diagnosis rather than ranking: an instrument that localizes where and why an agent'"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2606.29771","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-06-29T04:26:30Z","cross_cats_sorted":["cs.LG","q-fin.CP","q-fin.PM"],"title_canon_sha256":"301fe2e86ece1f086bea3a513cf37489295e416a221e094c42c20d4d1d7dc8c0","abstract_canon_sha256":"e4afc5cb9d4c6cc325cc5ccf18cb75c4d04bc2e6a8aa5e48bcb548750338285d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T02:17:34.555343Z","signature_b64":"bHV0STPLtSTdes3MGjwumKUNcrNY4XLR/4ENDZgMyRmMF6yo4FIWk9fO0chKhHSrubL0UHgsj0g7JLyqv0osCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6c7136420cec61989ed9ed02e6ff3b70be2dd491e438af8d0314d97819ab23df","last_reissued_at":"2026-06-30T02:17:34.554666Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T02:17:34.554666Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"CLQT: A Closed-Loop, Cost-Aware, Strategy-Consistent Benchmark for Diagnostic Evaluation of LLM Portfolio-Management Agents","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG","q-fin.CP","q-fin.PM"],"primary_cat":"cs.AI","authors_text":"Bo Qu, Mingguang Chen","submitted_at":"2026-06-29T04:26:30Z","abstract_excerpt":"LLM agents are increasingly cast as autonomous portfolio managers, and benchmarks have moved from financial question-answering to sequential trading. Yet most still rank agents by returns over a fixed window -- a weak proxy, since a period's return is dominated by the market path and apparent alpha can dissolve once look-ahead leakage is controlled. Such a ranking certifies neither sound reasoning, nor a consistent strategy, nor a durable edge. We introduce CLQT, which reframes closed-loop trading evaluation as diagnosis rather than ranking: an instrument that localizes where and why an agent'"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.29771","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2606.29771/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2606.29771","created_at":"2026-06-30T02:17:34.554764+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.29771v1","created_at":"2026-06-30T02:17:34.554764+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.29771","created_at":"2026-06-30T02:17:34.554764+00:00"},{"alias_kind":"pith_short_12","alias_value":"NRYTMQQM5RQZ","created_at":"2026-06-30T02:17:34.554764+00:00"},{"alias_kind":"pith_short_16","alias_value":"NRYTMQQM5RQZRHWZ","created_at":"2026-06-30T02:17:34.554764+00:00"},{"alias_kind":"pith_short_8","alias_value":"NRYTMQQM","created_at":"2026-06-30T02:17:34.554764+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC","json":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC.json","graph_json":"https://pith.science/api/pith-number/NRYTMQQM5RQZRHWZ5UBON7Z3OC/graph.json","events_json":"https://pith.science/api/pith-number/NRYTMQQM5RQZRHWZ5UBON7Z3OC/events.json","paper":"https://pith.science/paper/NRYTMQQM"},"agent_actions":{"view_html":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC","download_json":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC.json","view_paper":"https://pith.science/paper/NRYTMQQM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.29771&json=true","fetch_graph":"https://pith.science/api/pith-number/NRYTMQQM5RQZRHWZ5UBON7Z3OC/graph.json","fetch_events":"https://pith.science/api/pith-number/NRYTMQQM5RQZRHWZ5UBON7Z3OC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC/action/storage_attestation","attest_author":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC/action/author_attestation","sign_citation":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC/action/citation_signature","submit_replication":"https://pith.science/pith/NRYTMQQM5RQZRHWZ5UBON7Z3OC/action/replication_record"}},"created_at":"2026-06-30T02:17:34.554764+00:00","updated_at":"2026-06-30T02:17:34.554764+00:00"}