When Reasoning Meets Information Aggregation: A Case Study with Sports Narratives

TL;DR

This paper investigates how large language models reason about sports narratives, emphasizing the importance of accurate information aggregation, and introduces SportsGen for synthesizing game data to evaluate reasoning capabilities.

Contribution

It presents a new method, SportsGen, for generating sports narratives, and provides a comprehensive analysis of LLMs' reasoning performance on complex sports data.

Findings

Most models struggle with accurate score aggregation.

Open-source models like Llama-3 often hallucinate scores.

Narrative complexity affects reasoning effectiveness.

Abstract

Reasoning is most powerful when an LLM accurately aggregates relevant information. We examine the critical role of information aggregation in reasoning by requiring the LLM to analyze sports narratives. To succeed at this task, an LLM must infer points from actions, identify related entities, attribute points accurately to players and teams, and compile key statistics to draw conclusions. We conduct comprehensive experiments with real NBA basketball data and present SportsGen, a new method to synthesize game narratives. By synthesizing data, we can rigorously evaluate LLMs' reasoning capabilities under complex scenarios with varying narrative lengths and density of information. Our findings show that most models, including GPT-4o, often fail to accurately aggregate basketball scores due to frequent scoring patterns. Open-source models like Llama-3 further suffer from significant score hallucinations. Finally, the effectiveness of reasoning is influenced by narrative complexity, information density, and domain-specific terms, highlighting the challenges in analytical reasoning tasks.