Vendster: An Evaluation Arena to Investigate How Forecasting Tool Quality Affects LLM Agent Behaviors in Buyer-Vendor Negotiation
Source check: Source checked. Not independent scientific confirmation.
Finding: Forecast-quality metrics do not predict simulated negotiation outcomes; controlled under-prediction improves target attainment only for the two closed LLMs. This is the result-bearing preprint version dated 2026-10-28; peer-review status is not established by the reviewed source.
Technology: Four LLM families act as buyers/vendors using model-generated or controlled synthetic forecasts; the study varies forecast bias and interval width.
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Four LLM families act as buyers/vendors using model-generated or controlled synthetic forecasts; the study varies forecast bias and interval width.
- Test and result
- Forecast-quality metrics do not predict simulated negotiation outcomes; controlled under-prediction improves target attainment only for the two closed LLMs. The controlled bias/interval grid and observational forecast comparison use one M5-grounded simulated negotiation instance; no real sales or other elasticity is established.
- Limitation / still unknown
- not predict simulated negotiation outcomes; controlled under-prediction improves target attainment only for the two closed LLMs.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://arxiv.org/html/38292v1
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: not linked in this catalog; this does not mean no paper exists.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 2/5 · Validation 2/5 · Novelty 2/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.