Unlocking the Forecasting Economy: A Suite of Datasets for the Full Lifecycle of Prediction Market: [Experiments & Analysis]
Unlocking the Forecasting Economy: A Suite of Datasets for the Full Lifecycle of Prediction Market: [Experiments & Analysis]
Authors: Huaiyu Jia, Luofeng Zhou, Wentao Zhang, Lin William Cong, Siguang Li, Shuo Sun Date: 2026-04-22 Paper ID: openalex:2604.20421
Summary
This paper introduces a comprehensive, unified dataset suite for decentralized prediction markets, addressing the fragmentation of data across on-chain and off-chain sources. By integrating market metadata, high-granularity trading records, and oracle-resolution events, the authors provide a structured foundation for analyzing collective belief signals. The system supports reproducibility and extensibility, facilitating downstream research in event prediction, market calibration, and expectation reconstruction.
Key Contributions
- Introduces the first continuously maintained, unified dataset suite for the full lifecycle of decentralized prediction markets (Polymarket), covering October 2020 to March 2026.
- Develops a cross-source relational data system that integrates heterogeneous on-chain and off-chain data through identifier resolution and incremental synchronization.
- Demonstrates the dataset’s utility for time-series forecasting and analytical tasks via NBA outcome calibration and CPI expectation reconstruction case studies.
Open Questions & Future Work
Archivist Review
The paper focuses on the construction and utility of a large-scale, heterogeneous dataset for prediction markets. I have approved the dataset as a reusable resource for time-series and market research, and the open question regarding oracle risk as it presents a distinct, researchable problem in market dynamics. No novel algorithmic concepts were identified that transcend the domain-specific data collection effort.
Approved Open Questions
- Trading under oracle risk: Understanding oracle-risk market dynamics is critical for the robustness of decentralized oracle designs and for preventing information distortion in markets that depend on external settlement.
Datasets
Links
Metadata & Links
- url
- https://arxiv.org/abs/2604.20421
- paper_id
- 2604.20421
- paper_source
- openalex
- domain
- nlp
- tags
-
- architectures
-
- datasets
- polymarket-lifecycle-dataset
- concept_slugs
-
- dataset_slugs
- polymarket-lifecycle-dataset
- skill
- TimeSeriesSkill
- created_at
- 2026-04-25T05:38:06Z