{"record": {"aliases": ["FIFA World Cup 2026 as a Contamination-Free Benchmark for LLM Forecasting Agents: Four Models, a Bookmaker, and 104 Matches"], "artifacts": [{"id": "arxiv:2607.17765", "kind": "paper", "url": "https://arxiv.org/abs/2607.17765"}, {"id": "gh:graphuofm/fifa2026llm", "kind": "repo", "url": "https://github.com/graphuofm/FIFA2026LLM"}], "authors": ["Jiacheng Ding", "Cong Guo", "Jason Xu"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.17765", "source_url": "https://arxiv.org/abs/2607.17765"}, "categories": ["Agents & Tool Use", "Finance", "cs.LG"], "description": {"en": "WC2026-Agents evaluates autonomous LLM forecasting on 104 FIFA World Cup matches using probability forecasts, virtual wagers, and post-match reflections."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.17765", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.17765", "title": "WC2026-Agents"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["University of Memphis"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2607.17765", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2607.17765", "source_url": "https://openalex.org/W7169846290"}, "key": "claire-radar:2607.17765", "modality": null, "name": "WC2026-Agents", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_fifa-world-cup-2026-as-a-contamination-fre_e1616e27", "origin_source": "arxiv", "recognition_confidence": "0.7", "record_sha256": "336f14989c32bcfbeb9fba3bad52da890fdef73e0df9fceda046970cdc686503", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T17:28:54.793904Z", "source_url": "https://arxiv.org/abs/2607.17765"}, "publisher": null, "released": "2026-07-20", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.17765", "source_url": "https://arxiv.org/abs/2607.17765"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-17765", "source": "claire_radar", "source_benchmark_id": "2607.17765", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["FIFA World Cup 2026 as a Contamination-Free Benchmark for LLM Forecasting Agents: Four Models, a Bookmaker, and 104 Matches"], "annotation": "Unknown", "applicationDomains": ["Finance"], "area": "Agents & Tool Use", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The World Cup setting, frontier-model comparison, and contamination-resistant future-event design may generate substantial immediate interest.", "score": 80}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T17:28:54.793904Z", "inputHash": "0b60caeffdf5627aaed8da0fc5dfe1201eea0698d277e723d89533fba751ab00", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "We introduce WC2026-Agents, a benchmark and dataset", "canonicalNameSource": "abstract", "decisionReason": "The source formally names the benchmark, defines a repeated forecasting task, identifies Brier score and betting return on investment, and provides public data, code, and a reproducible evaluation suite.", "evaluationEvidence": {"reuse": "Data and code: https://github.com/graphuofm/FIFA2026LLM", "scoring": "none beats the market's Brier score", "task": "commit to a 1X2 (team-A win / draw / team-B win) distribution and a virtual 100-USD bet"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T17:28:54.793904Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "WC2026-Agents evaluates autonomous LLM forecasting on 104 FIFA World Cup matches using probability forecasts, virtual wagers, and post-match reflections.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce WC2026-Agents, a benchmark and dataset for evaluating large language models (LLMs) as autonomous forecasting agents on real, future events."}, "familyId": "bmf_079d787faa2c", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_fifa-world-cup-2026-as-a-contamination-fre_e1616e27", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Financial Services"], "links": {"code": "https://github.com/graphuofm/FIFA2026LLM", "data": null, "paper": "https://arxiv.org/abs/2607.17765", "pdf": "https://arxiv.org/pdf/2607.17765", "project": null, "report": "https://arxiv.org/abs/2607.17765"}, "metrics": [], "motivation": "We introduce WC2026-Agents, a benchmark and dataset for evaluating large language models (LLMs) as autonomous forecasting agents on real, future events.", "name": "WC2026-Agents", "oneLine": "WC2026-Agents evaluates autonomous LLM forecasting on 104 FIFA World Cup matches using probability forecasts, virtual wagers, and post-match reflections.", "paperTitle": "FIFA World Cup 2026 as a Contamination-Free Benchmark for LLM Forecasting Agents: Four Models, a Bookmaker, and 104 Matches", "primaryDomain": "Finance", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2607.17765", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "publishers": [{"name": "graphuofm", "organizationType": "community", "role": "benchmark-publisher", "sourceUrl": "https://github.com/graphuofm/FIFA2026LLM"}], "readiness": "Runnable", "recognitionConfidence": 0.7, "relation": "introduces", "releasedAt": "2026-07-20", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Jiacheng Ding", "Cong Guo", "Jason Xu"], "categories": ["cs.LG", "cs.AI", "cs.DB"], "id": "2607.17765", "title": "FIFA World Cup 2026 as a Contamination-Free Benchmark for LLM Forecasting Agents: Four Models, a Bookmaker, and 104 Matches", "type": "arxiv", "url": "https://arxiv.org/abs/2607.17765"}, "sourceUpdatedAt": "2026-07-20T10:00:11Z", "topics": ["cs.LG"], "venueAttempts": [], "whyItMatters": "It provides a future-event forecasting evaluation with an economically grounded betting-market baseline and measures calibration and decision quality."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
