pith. machine review for the scientific record. sign in

arxiv: 2604.20421 · v1 · submitted 2026-04-22 · 💻 cs.LG

Recognition: unknown

Unlocking the Forecasting Economy: A Suite of Datasets for the Full Lifecycle of Prediction Market: [Experiments \& Analysis]

Huaiyu Jia , Luofeng Zhou , Wentao Zhang , Lin William Cong , Siguang Li , Shuo Sun
Authors on Pith no claims yet

Pith reviewed 2026-05-10 01:25 UTC · model grok-4.3

classification 💻 cs.LG
keywords prediction marketsdecentralized financePolymarketdatasetoracle resolutiontrading datamarket lifecycleforecasting
0
0 comments X

The pith

A unified dataset suite now covers the full lifecycle of decentralized prediction markets on Polymarket from creation to settlement.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper establishes a continuously maintained collection of data for prediction markets on the Polymarket platform. It combines fragmented sources into a single relational system linking market details, every trade fill, and resolution outcomes by oracles. The dataset includes more than 770 thousand markets, 943 million trade records, and 2 million oracle events spanning from late 2020 into 2026. This integration solves problems of incomplete linkage and allows ongoing updates. Researchers can use it to analyze collective beliefs expressed through market prices and test forecasting accuracy in specific domains.

Core claim

The authors build a unified relational data system that integrates three canonical layers: market metadata, fill-level trading records, and oracle-resolution events, through identifier resolution, on-chain recovery, and incremental updates. The resulting dataset spans October 2020 to March 2026 and comprises more than 770 thousand market records, over 943 million fill records, and nearly 2 million oracle events. They demonstrate its utility through descriptive analyses and case studies on NBA outcome calibration and CPI expectation reconstruction.

What carries the argument

The unified relational data system integrating market metadata, trading records, and oracle events using identifier resolution and incremental updates.

Load-bearing premise

That the cross-source identifier resolution and incremental updates create complete and accurate links between off-chain metadata, on-chain trades, and oracle events with no major missing or mismatched records.

What would settle it

Finding a significant number of markets where the linked trade records or oracle events do not match the actual blockchain transactions or public resolutions would show the integration is incomplete.

Figures

Figures reproduced from arXiv: 2604.20421 by Huaiyu Jia, Lin William Cong, Luofeng Zhou, Shuo Sun, Siguang Li, Wentao Zhang.

Figure 1
Figure 1. Figure 1: Amount of total value locked (TVL), 30-day rolling [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustrative code of the polyMonitor pipeline for realtime acquisition and synchronization. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Series, events, markets, and tokens in Polymarket. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Overview of the full-lifecycle Polymarket data system. The pipeline integrates off-chain market metadata with [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Illustration of OrderFilled semantics in Polymarket. A single taker-side buy execution may correspond to multiple [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Quality gains from key ablation mechanisms. [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Primary topic analysis of Polymarket markets. (a) shows the keyword cloud of primary topics. (b) reports the top five primary topics [PITH_FULL_IMAGE:figures/full_fig_p008_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Normalized daily activity on Polymarket, measured [PITH_FULL_IMAGE:figures/full_fig_p008_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Post-anchor trading dynamics in wrong-case mar [PITH_FULL_IMAGE:figures/full_fig_p008_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Actual fee rate patterns across market-level trading activity and category groups. The left and middle panels plot the [PITH_FULL_IMAGE:figures/full_fig_p010_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Reliability diagram for size-weighted market prob [PITH_FULL_IMAGE:figures/full_fig_p010_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Time-series evolution of market-implied prob [PITH_FULL_IMAGE:figures/full_fig_p012_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Comparison of Polymarket-implied CPI, Cleveland nowcast, and BLS actual CPI across three target months under the size-weighted [PITH_FULL_IMAGE:figures/full_fig_p013_14.png] view at source ↗
read the original abstract

Prediction markets are markets for trading claims on future events, such as presidential elections, and their prices provide continuously updated signals of collective beliefs. In decentralized platforms such as Polymarket, the market lifecycle spans market creation, token registration, trading, oracle interaction, dispute, and final settlement, yet the corresponding data are fragmented across heterogeneous off-chain and on-chain sources. We present the first continuously maintained dataset suite for the full lifecycle of decentralized prediction markets, built on Polymarket. To address the challenges of large-scale cross-source integration, incomplete linkage, and continuous synchronization, we build a unified relational data system that integrates three canonical layers: market metadata, fill-level trading records, and oracle-resolution events, through identifier resolution, on-chain recovery, and incremental updates. The resulting dataset spans October 2020 to March 2026 and comprises more than 770 thousand market records, over 943 million fill records, and nearly 2 million oracle events. We describe the data model, collection pipeline, and consistency mechanisms that make the dataset reproducible and extensible, and we demonstrate its utility through descriptive analyses of market activity and two downstream case studies: NBA outcome calibration and CPI expectation reconstruction.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper claims to present the first continuously maintained dataset suite for the full lifecycle of decentralized prediction markets on Polymarket. It integrates three layers—market metadata, fill-level trading records, and oracle-resolution events—via identifier resolution, on-chain recovery, and incremental updates to produce a unified relational system. The resulting dataset spans October 2020–March 2026 and contains more than 770 thousand market records, over 943 million fill records, and nearly 2 million oracle events. The authors describe the data model, collection pipeline, and consistency mechanisms for reproducibility and extensibility, and illustrate utility through descriptive analyses of market activity plus two case studies on NBA outcome calibration and CPI expectation reconstruction.

Significance. If the integration mechanisms achieve complete and accurate linkage, the dataset would be a substantial resource for research on prediction markets, collective belief formation, and forecasting. Its scale, continuous maintenance, and coverage of the entire market lifecycle (creation through settlement) enable analyses that were previously limited by fragmented data sources. The emphasis on reproducibility, extensibility, and open case studies further enhances its value for downstream empirical work in the field.

major comments (1)
  1. Abstract: The central claim that the unified relational system produces a complete, error-free dataset via identifier resolution, on-chain recovery, and incremental updates is not supported by any quantitative validation. No success rates for cross-source identifier matching, mismatch frequencies, missing-record detection rates, or ground-truth completeness checks on any subset are reported, despite the dataset scale (770k markets, 943M fills, 2M oracle events) being presented as a direct outcome of these mechanisms. This is load-bearing for the paper's contribution, as the downstream case studies (NBA calibration, CPI reconstruction) and all activity statistics rest on the unverified assumption that integration artifacts do not distort the data.
minor comments (1)
  1. Abstract and title: The title contains the bracketed placeholder '[Experiments & Analysis]', which should be removed or clarified for a final manuscript. The abstract's scale figures ('more than 770 thousand', 'over 943 million', 'nearly 2 million') would benefit from exact counts and a dedicated table or section summarizing dataset statistics.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading and for highlighting the importance of quantitative validation for our integration claims. We agree this is a substantive point and will revise the manuscript accordingly to include explicit metrics and a dedicated validation section.

read point-by-point responses

  1. Referee: [—] Abstract: The central claim that the unified relational system produces a complete, error-free dataset via identifier resolution, on-chain recovery, and incremental updates is not supported by any quantitative validation. No success rates for cross-source identifier matching, mismatch frequencies, missing-record detection rates, or ground-truth completeness checks on any subset are reported, despite the dataset scale (770k markets, 943M fills, 2M oracle events) being presented as a direct outcome of these mechanisms. This is load-bearing for the paper's contribution, as the downstream case studies (NBA calibration, CPI reconstruction) and all activity statistics rest on the unverified assumption that integration artifacts do not distort the data.

    Authors: We acknowledge that the manuscript describes the integration mechanisms (identifier resolution, on-chain recovery, and incremental updates) but does not report quantitative validation metrics such as matching success rates, mismatch frequencies, or ground-truth completeness checks. This omission weakens the support for the completeness claims. In the revised version we will add a new subsection titled 'Validation of Integration Pipeline' (placed after the data model description) that reports: (i) success rates for cross-source identifier matching (e.g., percentage of markets successfully linked between off-chain metadata and on-chain records), (ii) observed mismatch frequencies and resolution procedures, (iii) missing-record detection rates via on-chain recovery, and (iv) results of ground-truth completeness audits on sampled subsets (e.g., manual verification of 500–1,000 randomly selected markets and oracle events). We will also discuss any residual limitations and their potential impact on the NBA and CPI case studies. These additions will make the load-bearing assumptions explicit and verifiable. revision: yes

Circularity Check

0 steps flagged

No circularity: data construction paper with no derivations or fitted predictions

full rationale

The paper presents a dataset construction pipeline for Polymarket prediction markets, integrating metadata, trades, and oracle events via identifier resolution and incremental updates. No mathematical derivations, equations, fitted parameters, or 'predictions' are claimed that could reduce to the input data by construction. The output is the dataset and descriptive analyses, not a result forced by self-definition or self-citation chains. The work is self-contained as a data engineering effort without load-bearing theoretical steps.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the domain assumption that Polymarket's heterogeneous sources can be reliably linked via identifiers and that oracle events accurately reflect resolutions; no free parameters or invented entities are introduced.

axioms (1)
  • domain assumption Off-chain and on-chain data sources share consistent identifiers that allow accurate linkage without substantial loss or duplication.
    Invoked in the description of the unified relational data system and identifier resolution process.

pith-pipeline@v0.9.0 · 5529 in / 1042 out tokens · 30935 ms · 2026-05-10T01:25:41.328469+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

41 extracted references · 7 canonical work pages · 1 internal anchor

  1. [1]

    Kalshi and the rise of macro markets

    Anthony M Diercks, Jared Dean Katz, and Jonathan H Wright. Kalshi and the rise of macro markets. Technical report, National Bureau of Economic Research, 2026

    2026

  2. [2]

    Price discovery and trading in modern prediction markets.A vailable at SSRN 5331995, 2026

    Hunter Ng, Lin Peng, Yubo Tao, and Dexin Zhou. Price discovery and trading in modern prediction markets.A vailable at SSRN 5331995, 2026

    2026

  3. [3]

    Financial and informa- tional integration through oracle networks

    Lin William Cong, Eswar S Prasad, and Daniel Rabetti. Financial and informa- tional integration through oracle networks. Technical report, National Bureau of Economic Research, 2025

    2025

  4. [4]

    Ex-graph: A pioneering dataset bridging ethereum and x.arXiv preprint arXiv:2310.01015, 2023

    Qian Wang, Zhen Zhang, Zemin Liu, Shengliang Lu, Bingqiao Luo, and Bingsheng He. Ex-graph: A pioneering dataset bridging ethereum and x.arXiv preprint arXiv:2310.01015, 2023

    work page arXiv 2023

  5. [5]

    Nft1000: A cross-modal dataset for non- fungible token retrieval

    Shuxun Wang, Yunfei Lei, Ziqi Zhang, Wei Liu, Haowei Liu, Li Yang, Bing Li, Wenjuan Li, Jin Gao, and Weiming Hu. Nft1000: A cross-modal dataset for non- fungible token retrieval. InProceedings of the 32nd ACM International Conference on Multimedia, pages 2214–2222, 2024

    2024

  6. [6]

    Token spammers, rug pulls, and sniper bots: An analysis of the ecosystem of tokens in ethereum and in the binance smart chain ({ { { { {BNB} } } } })

    Federico Cernera, Massimo La Morgia, Alessandro Mei, and Francesco Sassi. Token spammers, rug pulls, and sniper bots: An analysis of the ecosystem of tokens in ethereum and in the binance smart chain ({ { { { {BNB} } } } }). In32nd USENIX security symposium (USENIX security 23), pages 3349–3366, 2023

    2023

  7. [7]

    A midsummer meme’s dream: Investigating market manipulations in the meme coin ecosystem.arXiv preprint arXiv:2507.01963, 2025

    Alberto Maria Mongardini and Alessandro Mei. A midsummer meme’s dream: Investigating market manipulations in the meme coin ecosystem.arXiv preprint arXiv:2507.01963, 2025

    work page arXiv 2025

  8. [8]

    SoK : Market microstructure for decentralized prediction markets ( DePMs )

    Nahid Rahman, Joseph Al-Chami, and Jeremy Clark. Sok: Market microstructure for decentralized prediction markets (depms).arXiv preprint arXiv:2510.15612, 2025

    work page arXiv 2025

  9. [9]

    Are betting markets better than polling in predicting political elections?arXiv preprint arXiv:2507.08921, 2025

    Laurie E Cutting, Sarah S Hughes-Berheim, Paul M Johnson, Hiba Baroud, and Brett Goldstein. Are betting markets better than polling in predicting political elections?arXiv preprint arXiv:2507.08921, 2025

    work page arXiv 2025

  10. [10]

    Political uncertainty and credit risk: The role of event markets in forecasting ukraine’s sovereign spreads.A vailable at SSRN 5163719, 2025

    Zachary McGurk and Mary J Becker. Political uncertainty and credit risk: The role of event markets in forecasting ukraine’s sovereign spreads.A vailable at SSRN 5163719, 2025

    2025

  11. [11]

    Under pressure? central bank independence meets blockchain prediction markets

    Barry Eichengreen, Ganesh Viswanath-Natraj, Junxuan Wang, and Zijie Wang. Under pressure? central bank independence meets blockchain prediction markets. Central Bank Independence Meets Blockchain Prediction Markets (July 26, 2025), 2025

    2025

  12. [12]

    arXiv preprint arXiv:2508.03474 , year =

    Oriol Saguillo, Vahid Ghafouri, Lucianna Kiffer, and Guillermo Suarez-Tangil. Unravelling the probabilistic forest: Arbitrage in prediction markets.arXiv preprint arXiv:2508.03474, 2025

    work page arXiv 2025

  13. [13]

    Semantic trading: Agentic ai for clustering and relationship discovery in prediction markets.arXiv preprint arXiv:2512.02436, 2025

    Agostino Capponi, Alfio Gliozzo, and Brian Zhu. Semantic trading: Agentic ai for clustering and relationship discovery in prediction markets.arXiv preprint arXiv:2512.02436, 2025

    work page arXiv 2025

  14. [14]

    Onchain reputation: Introducing dynamic incentives into distributed networks.A vailable at SSRN 4540862, 2023

    Lin William Cong and Luofeng Zhou. Onchain reputation: Introducing dynamic incentives into distributed networks.A vailable at SSRN 4540862, 2023

    2023

  15. [15]

    A primer on oracle economics.Journal of Corporate Finance, 94:102800, 2025

    Lin William Cong, Liam Fox, Siguang Li, and Luofeng Zhou. A primer on oracle economics.Journal of Corporate Finance, 94:102800, 2025

    2025

  16. [16]

    A brief introduction to blockchain economics

    Long Chen, Lin William Cong, and Yizhou Xiao. A brief introduction to blockchain economics. InInformation for efficient decision making: Big data, blockchain and relevance, pages 1–40. World Scientific, 2021

    2021

  17. [17]

    Diablo: A benchmark suite for blockchains

    Vincent Gramoli, Rachid Guerraoui, Andrei Lebedev, Chris Natoli, and Gauthier Voron. Diablo: A benchmark suite for blockchains. InProceedings of the Eighteenth European Conference on Computer Systems, pages 540–556, 2023

    2023

  18. [18]

    Multi-chain graphs of graphs: A new approach to analyzing blockchain datasets.Advances in Neural Information Processing Systems, 37:28490–28514, 2024

    Bingqiao Luo, Zhen Zhang, Qian Wang, and Bingsheng He. Multi-chain graphs of graphs: A new approach to analyzing blockchain datasets.Advances in Neural Information Processing Systems, 37:28490–28514, 2024

    2024

  19. [19]

    Zipzap: Efficient training of language models for large-scale fraud detection on blockchain

    Sihao Hu, Tiansheng Huang, Ka-Ho Chow, Wenqi Wei, Yanzhao Wu, and Ling Liu. Zipzap: Efficient training of language models for large-scale fraud detection on blockchain. InProceedings of the ACM Web Conference 2024, pages 2807–2816, 2024

    2024

  20. [20]

    Graph neural networks for bridge swap link prediction in uniswap v3

    Qingran Zhou, Eric Liu, and Alessio Brini. Graph neural networks for bridge swap link prediction in uniswap v3. InProceedings of the 6th ACM International Conference on AI in Finance, pages 247–255, 2025

    2025

  21. [21]

    Money never sleeps: Maximizing liquidity mining yields in decentralized finance

    Wangze Ni, Zhao Yiwei, Weijie Sun, Lei Chen, Peng Cheng, Chen Jason Zhang, and Xuemin Lin. Money never sleeps: Maximizing liquidity mining yields in decentralized finance. InProceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, pages 2248–2259, 2024

    2024

  22. [22]

    Sok: Decentralized finance (defi) attacks

    Liyi Zhou, Xihan Xiong, Jens Ernstberger, Stefanos Chaliasos, Zhipeng Wang, Ye Wang, Kaihua Qin, Roger Wattenhofer, Dawn Song, and Arthur Gervais. Sok: Decentralized finance (defi) attacks. In2023 IEEE Symposium on Security and Privacy (SP), pages 2444–2461. IEEE, 2023

    2023

  23. [23]

    Security perceptions of users in stablecoins: Advantages and risks within the cryptocurrency ecosystem

    Maggie Yongqi Guan, Yaman Yu, Tanusree Sharma, Molly Zhuangtong Huang, Kaihua Qin, Yang Wang, and Kanye Ye Wang. Security perceptions of users in stablecoins: Advantages and risks within the cryptocurrency ecosystem. In2025 IEEE Symposium on Security and Privacy (SP), pages 2753–2771. IEEE, 2025

    2025

  24. [24]

    The Price of Interoperability: Exploring Cross-Chain Bridges and Their Economic Consequences

    Yiyue Cao, Mingzhe Zheng, Lin William Cong, Siguang Li, and Xuechao Wang. The price of interoperability: Exploring cross-chain bridges and their economic consequences.arXiv preprint arXiv:2604.03083, 2026

    work page internal anchor Pith review Pith/arXiv arXiv 2026

  25. [25]

    Stop pulling my rug: Exposing rug pull risks in crypto token to investors

    Yuanhang Zhou, Jingxuan Sun, Fuchen Ma, Yuanliang Chen, Zhen Yan, and Yu Jiang. Stop pulling my rug: Exposing rug pull risks in crypto token to investors. InProceedings of the 46th International Conference on Software Engi- neering: Software Engineering in Practice, pages 228–239, 2024

    2024

  26. [26]

    Sok: A taxonomic analysis of defi rug pulls: Types, dataset, and tool assessment.Proceedings of the ACM on Software Engineering, 2(ISSTA):550–572, 2025

    Dianxiang Sun, Wei Ma, Liming Nie, and Yang Liu. Sok: A taxonomic analysis of defi rug pulls: Types, dataset, and tool assessment.Proceedings of the ACM on Software Engineering, 2(ISSTA):550–572, 2025

    2025

  27. [27]

    Characterizing ethereum address poisoning attack

    Shixuan Guan and Kai Li. Characterizing ethereum address poisoning attack. In Proceedings of the 2024 on ACM SIGSAC Conference on Computer and Communi- cations Security, pages 986–1000, 2024

    2024

  28. [28]

    Polymarket Documentation

    Polymarket. Polymarket Documentation. https://docs.polymarket.com/, 2026. Accessed: 2026-04-07

    2026

  29. [29]

    Polygon pos chain explorer, 2026

    Polygonscan. Polygon pos chain explorer, 2026. Accessed: 2026-04-07

    2026

  30. [30]

    Polymarket CTF exchange

    Polymarket. Polymarket CTF exchange. https://github.com/polymarket/ctf- exchange, 2022. Accessed: 2026-04-13

    2022

  31. [31]

    Negrisk CTF adapter: Multi-outcome market contracts

    Polymarket. Negrisk CTF adapter: Multi-outcome market contracts. https: //github.com/polymarket/neg-risk-ctf-adapter, 2023. Accessed: 2026-04-13; Re- lease v2.0.0

    2023

  32. [32]

    UMA: A decentralized optimistic oracle

    Risk Labs Foundation. UMA: A decentralized optimistic oracle. https://uma.xyz/,

  33. [33]

    Accessed: 2026-04-13

    2026

  34. [34]

    Polymarket. Fees. https://docs.polymarket.com/trading/fees, 2026. Polymarket Documentation, accessed April 10, 2026

    2026

  35. [35]

    Polygon transaction record for the first observed non-zero fee trade on polymarket, 2026

    PolygonScan. Polygon transaction record for the first observed non-zero fee trade on polymarket, 2026. Transaction hash: 0xb05478...f8fd88c. Accessed: 2026-04-10

    2026

  36. [36]

    Live sports prediction markets

    Polymarket. Live sports prediction markets. https://polymarket.com/sports/live,

  37. [37]

    Accessed: 2026-04-13; Real-time market data subject to change

    2026

  38. [38]

    The isotonic regression problem and its dual.Journal of the American Statistical Association, 67(337):140–147, 1972

    Richard E Barlow and Hugh D Brunk. The isotonic regression problem and its dual.Journal of the American Statistical Association, 67(337):140–147, 1972

    1972

  39. [39]

    Consumer price index (cpi)

    International Monetary Fund. Consumer price index (cpi). https://data.imf.org/ en/datasets/IMF.STA:CPI, 2026. IMF Data dataset page. Accessed: 2026-04-13

    2026

  40. [40]

    Inflation nowcasting

    Federal Reserve Bank of Cleveland. Inflation nowcasting. https://www. clevelandfed.org/indicators-and-data/inflation-nowcasting, 2026. Center for Inflation Research; updated each business day. Accessed: 2026-04-07

    2026

  41. [41]

    Bureau of Labor Statistics

    U.S. Bureau of Labor Statistics. Schedule of releases for the consumer price index. https://www.bls.gov/schedule/news_release/cpi.htm, 2026. Accessed: 2026-04-13. 13

    2026