ForesightFlow Datasets

Public datasets released alongside ForesightFlow research.

Index

Dataset	Description	License	Size	Tag
coordination-traces-100	500 LLM reasoning traces (100 markets × 5 coordination configs) from the Phase 0.5 shakedown of "Coordination as an Architectural Layer for LLM-Based Multi-Agent Systems"	CC-BY 4.0	8.8 MB	coordination-traces-100-v1
ffic-inventory	8 publicly documented Polymarket insider-trading cases mapped to 24 on-chain market identifiers, released alongside the paper "ForesightFlow: Real-Time Detection of Informed Trading in Decentralized Prediction Markets"	CC-BY 4.0	21 KB	ffic-inventory-v1
polymarket-deadline-ils	Population-scale Deadline-ILS (ILS^dl) scores for 88 Polymarket markets across military/geopolitical, regulatory, and corporate categories. Includes LLM-recovered event dates (T_event), bootstrap CIs, and full 2,375-market attrition chain. Snapshot: 2020–2026.	CC-BY 4.0	311 KB	polymarket-deadline-ils-v3
polymarket-tnews-tevent-recovery	Curated public-event and news-arrival timestamps for 2,052 resolved Polymarket markets across three methodological tiers: UMA Oracle proposer evidence (Tier 1, n=12), GDELT proxy (Tier 2, n=1,993), and LLM-assisted multi-source verification (Tier 3, n=47). Snapshot: 2022-12–2026-04.	CC-BY 4.0	~6 MB	polymarket-tnews-tevent-recovery-v1
polymarket-hazard-rates	Per-category exponential hazard fits (MLE λ̂, 95% CI, KS test) for the time-to-event distribution on Polymarket deadline-resolved contracts. Baseline survival function for the ILS-dl framework. Categories: military_geopolitics (n=18, λ=0.241, adequate), corporate_disclosure (preliminary), regulatory_decision (rejected), esports (n/a).	CC-BY 4.0	<1 MB	polymarket-hazard-rates-v1
polymarket-ils-corpus	Population-scale ILS corpus for 4,801 resolved Polymarket markets. Anchor: t_resolve − 24h proxy (4,796 markets, 99.9%) or recovered T_event (5 markets). Multi-window variants, scope flags, HHI. anchor_type column distinguishes proxy from event-anchored records. 2,548 clean-scope markets.	CC-BY 4.0	~4 MB	polymarket-ils-corpus-v1
polymarket-resolution-typology	Three-class classification of 911,237 Polymarket markets by resolution mechanism (deadline_resolved / event_resolved / unclassifiable), with category labels and volume metadata. Snapshot: 2020–2026-04-27.	CC-BY 4.0	242 MB (LFS)	polymarket-resolution-typology-v1
pmxt-stylized-facts-v1	Per-market stylized-fact measurements (SF1–SF9) for 13,314 resolved Polymarket binary-event markets, week 2026-04-21 to 2026-04-27. Empirical foundation for Nechepurenko (2026) Paper 1 (event-linked perpetuals).	CC-BY 4.0	1.6 MB	pmxt-stylized-facts-v1
pmxt-counterfactual-replay-v1	Counterfactual simulation results (E2 margin recalibration + E3 resolution-zone protocol comparison) for 13,000+ resolved Polymarket binary-event markets. Per-(engine/mechanic, leverage, class) liquidation rates, bad-debt frequencies, drawdown, and PnL. Companion to pmxt-stylized-facts-v1 and Nechepurenko (2026) Paper 1.	CC-BY 4.0	19 KB	pmxt-counterfactual-replay-v1
pmxt-behavioral-clusters-v1	Fill-side behavioral clusters, feature tiers, and per-market microstructure signatures from 13.4M OrderFilled events on Polymarket CTFExchange (43,116 markets, 77,203 addresses, 2026-04-21 to 2026-04-27). k-means k=5 archetypes, 6 reviewer-defensible feature tiers, bilateral Spearman analysis with BH-FDR correction. Companion to Nechepurenko (2026) Paper 4.	CC-BY 4.0	17.8 MB (LFS)	pmxt-behavioral-clusters-v1

---

PMXT Bundle Family (Event-Linked Perpetuals)

Datasets released as the empirical foundation for the four-paper Event-Linked Perpetuals programme by Maksym Nechepurenko (Devnull Research). See each bundle's README for schema, methodology, and citation details.

Bundle	Folder	DOI	Status
Bundle 1	pmxt-stylized-facts-v1/	10.5281/zenodo.20107449	Released
Bundle 2	pmxt-counterfactual-replay-v1/	10.5281/zenodo.20108387	Released
Bundle 3	pmxt-behavioral-clusters-v1/	TBD (Zenodo forthcoming)	Released

---

Adding a new dataset

Create a subdirectory under the repo root. Each dataset directory must contain:

• README.md — description, schema, quick-start
• DATASHEET.md — Gebru et al. (2021) datasheet
• CITATION.cff — citation metadata
• LICENSE — dataset license (CC-BY 4.0 recommended)
• data/ — data files

Update this top-level README index, then tag as <dataset-name>-v1.

---

Cite this work

If you use these datasets, please cite the papers they accompany:

Information Leakage at Population Scale

@misc{nechepurenko2026population-leakage,
  title  = {Information Leakage at Population Scale: An Evaluation of the {Polymarket} Insider-Relevant Subpopulation},
  author = {Nechepurenko, Maksym},
  year   = {2026},
  doi    = {10.48550/arXiv.2605.00459},
  url    = {https://arxiv.org/abs/2605.00459},
  note   = {SSRN Working Paper 6686819}
}

ForesightFlow: An Information Leakage Score Framework for Prediction Markets

@misc{nechepurenko2026ils-framework,
  title  = {{ForesightFlow}: An Information Leakage Score Framework for Prediction Markets},
  author = {Nechepurenko, Maksym},
  year   = {2026},
  doi    = {10.48550/arXiv.2605.00493},
  url    = {https://arxiv.org/abs/2605.00493},
  note   = {SSRN Working Paper 6687361}
}

Empirical Evaluation of Deadline-Resolved Information Leakage on Documented Polymarket Insider Cases

@misc{nechepurenko2026deadline-leakage,
  title  = {Empirical Evaluation of Deadline-Resolved Information Leakage on Documented {Polymarket} Insider Cases},
  author = {Nechepurenko, Maksym},
  year   = {2026},
  doi    = {10.48550/arXiv.2605.02286},
  url    = {https://arxiv.org/abs/2605.02286},
  note   = {SSRN Working Paper 6687398}
}

Per-Market Information Leakage and Order-Flow Skill

@misc{nechepurenko2026per-market-ils,
  title  = {Per-Market Information Leakage and Order-Flow Skill: Two Methodological Lenses on Informed Trading in Decentralized Prediction Markets},
  author = {Nechepurenko, Maksym},
  year   = {2026},
  doi    = {10.48550/arXiv.2605.02287},
  url    = {https://arxiv.org/abs/2605.02287},
  note   = {SSRN Working Paper 6687441}
}

Coordination as an Architectural Layer for LLM-Based Multi-Agent Systems

@misc{nechepurenko2026coordination,
  title  = {Coordination as an Architectural Layer for LLM-Based Multi-Agent Systems: An Information-Controlled Empirical Study on Prediction Markets},
  author = {Nechepurenko, Maksym and Shuvalov, Pavel},
  year   = {2026},
  url    = {https://papers.ssrn.com/abstract=6687518},
  note   = {SSRN Working Paper 6687518}
}

Full preprint: https://foresightflow.org/publications/coordination-architectural-layer.

Resolution-Aware Perpetual Futures on Binary Prediction Markets

@misc{nechepurenko2026elp,
  title  = {Resolution-Aware Perpetual Futures on Binary Prediction Markets: An Empirical Risk-Design Framework Using Polymarket Data},
  author = {Nechepurenko, Maksym},
  year   = {2026},
  note   = {Working paper. Companion datasets: pmxt-stylized-facts-v1
            (DOI: 10.5281/zenodo.20107449), pmxt-counterfactual-replay-v1
            (DOI: 10.5281/zenodo.20108387).}
}

Fill-Side Non-Retail Trading on Polymarket

@misc{nechepurenko2026pmxt_clusters_paper,
  title  = {Fill-Side Non-Retail Trading on Polymarket: An Empirical Study of Behavioral
            Tiers and Microstructure Signatures Under Quote-Attribution Constraints},
  author = {Nechepurenko, Maksym},
  year   = {2026},
  note   = {Working paper. Companion dataset: pmxt-behavioral-clusters-v1
            (DOI: 10.5281/zenodo.XXXXXXXX).}
}