Documentation · 07 of 07

Chinese Competitive Debate: A Dataset and Benchmark with Professional Adjudication

Zongrui Yang, Haoyuan Li, Zhongsheng Wang, Zhirui Zeng, Pengqian Han, Yi Zhou, Yuting Wang, Jiamou Liu · Under review as a conference paper at ICLR 2027 · preprint arXiv:2609.21637

Download PDFarXiv 2609.21637Cite
01 · Abstract

Evaluating debate requires interpreting contributions in context and relating them to local assessments and overall outcomes. We introduce the Chinese Competitive Debate Dataset (CCDD), linking competition transcripts to original adjudication under a predefined rubric. CCDD contains 148 matches, 2,690 stage–side scoring instances, and 20,542 turns, with reviewed transcripts, aligned individual judge scores, match decisions, best-debater ballots, and commentary. Linked appeal records preserve reassessments. Three benchmark tasks — winner prediction, stage-score prediction, and best-debater prediction — demonstrate the use of this supervision. Zero-shot results reveal challenges in reproducing stage assessments, while judge-agreement analyses and structural references contextualize performance. CCDD supports research on contextual debate evaluation and judge disagreement.

02 · Contributions

Original professional adjudication records

Authentic Chinese competitive debate linked to individual stage scores, match decisions, best-debater votes and commentary collected under a predefined rubric.

Reviewed and aligned debate data

Transcripts and assessments organised with explicit stage and side associations; individual judgments and linked appeal adjudications preserved; provenance of original and generated materials documented.

A multilevel adjudication benchmark

Three prediction tasks with initial zero-shot results, judge-agreement analyses and structural references for reproducible evaluation and interpretation.

03 · Chinese competitive debate resources
Resource and scopeTextual organisation and focusAnnotations or assessment targetsAnnotation provenance
ORCHID (2023) · 1,218 debatesDebate transcripts and stance-specific summariesTarget-independent stance; argumentative dialogue summarisationCurated competition transcripts with task-specific annotations and summaries
CEDAR (2026) · 600 debatesComplete opposing-side transcripts with content, role and timing informationClaims, stances, evidence, rhetorical figures, modal words, debate resultsCollected competition records annotated for computational argumentation
Conch (2026) · 3 debates across languagesSessions, turns, argument blocks and linked clash structuresClash points, disagreements, viewpoints and strategiesLLM-assisted structural analysis with expert annotation and validation; designed for visual analysis
Chen et al. (2026) · 94 debatesContributions in open discussion and initial presentationsSupport, response and language ratings; debate outcomesLLM-generated quality ratings; competition outcomes; separate human pairwise validation
DEFINED (2026) · statement-levelCompetition statements and augmented materialsEight-dimensional creativity-related assessment and overall scoringExpert-scored authentic statements with mixed-granularity supervision and augmentation
CCDD · 148 retained matchesMatches, scheduled stages and turns, linked to evaluated sides and rolesStage scores, match decisions, best-debater ballots and judge commentaryIndividual competition-time judgments under a predefined rubric; linked appeal records; separately identified generated supplements

Counts denote debates; Conch includes one Chinese and two English debates. DEFINED is described by its statement-level assessment scope.

04 · Scope and limitations

CCDD covers the collected Chinese-language competitions and their formats; repeated participants, teams, judges and motions introduce potential dependencies, and filtering for usable scoring records may affect corpus composition. Manual review supports record integrity but does not estimate residual transcription or segmentation error. The experiments measure text-based zero-shot agreement with recorded adjudication; judges observed audiovisual performance and assessed the completed match, whereas stage-scoring models receive text and preceding context, which limits direct comparison.

05 · AI use statement

Generative AI tools were used to repair ASR transcripts under specified constraints, to process data and to summarise judges' rationales as supplementary annotations; processed transcripts and generated supplements underwent the review procedures described in the paper. AI tools also assisted with language polishing. The authors remain responsible for all claims, numerical results, citations and materials.

← PrivacyHome →