Metraipolis › archive › dc_7a370bfaff4b

FinalityBench: An Effect-Level Benchmark for Agent Decisions Under Delayed and Conflicting Financial Finality — arXiv:2609.04706

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.04706
License (as archived): CC0-1.0
Archived: 2026-10-16T22:20:51.286Z
SHA-256 of the archived text: fac4d93ea5cbe9559fbb7d4fd9b7ca99cf1e9a581418d25cd9ef0f5d1159a1f8
Stable URL of this page: https://metraipolis.com/documento/dc_7a370bfaff4b

How to cite

FinalityBench: An Effect-Level Benchmark for Agent Decisions Under Delayed and Conflicting Financial Finality — arXiv:2609.04706 — arXiv (en), https://arxiv.org/abs/2609.04706 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_7a370bfaff4b (Metraipolis archive dc_7a370bfaff4b, sha256 fac4d93ea5cbe9559fbb7d4fd9b7ca99cf1e9a581418d25cd9ef0f5d1159a1f8)

BibTeX

@misc{dc7a370bfaff4b,
  title = {FinalityBench: An Effect-Level Benchmark for Agent Decisions Under Delayed and Conflicting Financial Finality — arXiv:2609.04706},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.04706},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_7a370bfaff4b; sha256 fac4d93ea5cbe9559fbb7d4fd9b7ca99cf1e9a581418d25cd9ef0f5d1159a1f8},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.04706. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A merchant's payment processor, ledger, ERP and bank feed are updated by messages that get delayed, duplicated, dropped and reordered, so for minutes at a time the four hold contradictory beliefs about the same order. An agent resolving the exception must decide whether to ship goods, re-submit a capture, refund or wait, knowing some of those cannot be undone. We present FinalityBench, an executable benchmark for that decision. It keeps a hidden canonical event log and derives each system's view from a separately faulted delivery stream, so disagreement follows from specified fault semantics rather than being authored. Grading is on executed monetary effects: an episode is scored by the merchant's terminal economic position, relative to a privileged reference told when the pending capture resolves. The corpus of 321 tasks includes 45 twin pairs (90 tasks): tasks whose four system views are identical at the decision instant, whose authoritative probes both return unknown, and whose eventual correct dispositions differ. That snapshot indistinguishability is checked under every evaluation seed rather than assumed; equivalence over all interaction traces is not claimed. Over 14,445 graded episodes from nine programmatic policies, ranking by single-task accuracy and by paired loss disagree in 7 places: a ship-on-first-sign policy is second-best by accuracy at 65.7% and worst in the suite by paired loss, because it cannot tell the two members apart. A runtime gating irreversible actions on an authoritative finality probe reaches 85.4% and, unlike every polling policy, loses nothing to pass^5; its residual loss is almost entirely one archetype, which prices finality information directly. Language models reach the same exact rate as the hand-written gate on a stratified subset, lose about twice as much money, and discover the finality-gating strategy without being told it.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-07 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2609.04706 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T22:18:11.381Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).