Population-Scale Deadline-ILS Scores for Polymarket
Deadline Information Leakage Scores (ILS^dl) for 88 Polymarket markets across military/geopolitical, regulatory, and corporate categories. Includes LLM-recovered event dates (T_event), bootstrap confidence intervals, and the full 2,375-market attrition chain. Snapshot: 2020–2026.
This dataset accompanies Empirical Evaluation of Deadline-Resolved Information Leakage on Documented Polymarket Insider Cases (Nechepurenko, 2026). It provides ILS^dl scores for 88 Polymarket markets together with the full attrition chain documenting how 12,708 candidate markets were filtered to the final scored set.
ILS^dl Formula
ILS^dl(M) = (p(T_event⁻) − p_open) / (p_resolve − p_open)
p(T_event⁻) is the market price just before the real-world event; p_open is the opening price; p_resolve ∈ {0, 1} is the binary resolution outcome. A score near 1 indicates the price fully moved before the event — a strong informed-trading signal.
Contents
| File | Format | Rows | Description |
|---|---|---|---|
data/population_ils_dl.parquet | Parquet | 2,375 | Full population with T_event, ILS^dl, exclusion chain |
data/population_ils_dl.csv | CSV | 2,375 | Same data in flat CSV |
Scope
- Source markets: 12,708 (military_geopolitics + regulatory_decision + corporate_disclosure, volume ≥ $50K USDC)
- After pre-filter: 2,375 non-unclassifiable markets
- T_event recovered: 442 markets (Claude Haiku + web search)
- ILS^dl computed: 88 markets (require T_event confidence ≥ 0.7 and historical CLOB coverage)
- Bootstrap CI (B=500): 78 of 88 markets
Quick start
import pandas as pd
df = pd.read_parquet("data/population_ils_dl.parquet")
scored = df[df["ils_dl"].notna()] # 88 markets with ILS^dl
print(scored[["market_id", "ils_dl", "ci_low", "ci_high"]].head())
Distribution (88 markets)
| Metric | Value |
|---|---|
| Mean | −0.515 |
| Median | −0.395 |
| Std | 1.043 |
| Range | −7.750 to 0.994 |
Citation
@misc{nechepurenko2026deadline-ils-dataset,
title = {polymarket-deadline-ils: Population-Scale Deadline-ILS Scores for Polymarket},
author = {Nechepurenko, Maksym},
year = {2026},
publisher = {ForesightFlow / Devnull FZCO},
url = {https://github.com/ForesightFlow/datasets/tree/main/polymarket-deadline-ils},
note = {Version paper3a-v3, CC-BY-4.0. Accompanies: Empirical Evaluation of Deadline-Resolved Information Leakage on Documented Polymarket Insider Cases}
}