arxiv: 2605.02286 · v1 · submitted 2026-05-04 · 💱 q-fin.TR · q-fin.GN

Recognition: unknown

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

Maksym Nechepurenko

Pith reviewed 2026-05-08 01:30 UTC · model grok-4.3

classification 💱 q-fin.TR q-fin.GN

keywords information leakageprediction marketsPolymarketinsider tradingdeadline-resolved contractshazard rate estimation

0 comments

The pith

The deadline-Information Leakage Score produces a positive value on the largest Iran conflict contract while a proxy yields negative, distinguishing signal from artifact.

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

This paper empirically tests an extension of the information leakage score called ILS-dl for prediction market contracts that resolve on deadlines rather than events. It anchors the test in the largest documented cluster of Polymarket insider cases from the 2026 U.S.-Iran conflict. The central computation shows that the new score shifts from -0.331 under a resolution-anchored proxy to +0.113 when using article-derived public-event timestamps. Hazard rates receive exponential fits for military-geopolitics markets with a half-life of 2.3 days, while short time windows give zero leakage. A sympathetic reader would care because the method offers a concrete way to measure whether pre-deadline trading contains timing information beyond what the final resolution date alone would imply.

Core claim

On the largest applicable contract in the Iran cluster, the article-derived public-event timestamp yields ILS-dl = +0.113 versus a resolution-anchored proxy value of -0.331, a 0.444 shift in magnitude on opposite sides of zero. This demonstrates that the deadline extension distinguishes signal from proxy artefact. Hazard-rate estimation produces an adequate exponential fit for military-geopolitics markets and a preliminary fit for corporate-disclosure markets, while the regulatory-decision category is rejected as bimodal. Pre-event drift is mild and short-window variants are exactly zero. Cross-market wallet analysis identifies 332 wallets active in both major Iran-cluster markets.

What carries the argument

The deadline-Information Leakage Score (ILS-dl), which extends the original ILS to deadline-resolved contracts by anchoring calculations to public-event timestamps instead of resolution dates.

If this is right

The ILS-dl extension applies directly to other deadline-based prediction market contracts to quantify information leakage.
Military-geopolitics markets exhibit exponential hazard rates with a 2.3-day half-life.
Short time windows of 30 minutes and 2 hours around the public event show exactly zero leakage.
Hundreds of wallets trade across related contracts in the same cluster, though data covers only the resolution-settlement window.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If ILS-dl reliably separates signal, it could support automated screening of public prediction market data for timing anomalies.
Linking the score to wallet overlap patterns might surface coordinated trading across related deadline contracts.
The approach could be reapplied to smaller clusters in the same inventory to test whether the sign-flip result generalizes.

Load-bearing premise

The ForesightFlow Insider Cases inventory correctly identifies true insider trading instances and article-derived public-event timestamps accurately represent the moment information became public.

What would settle it

Recomputing ILS-dl for the same Iran contract with an independent public-event timestamp or an alternative set of identified insider trades that produces a value near -0.331 or a much smaller shift across zero.

read the original abstract

This paper reports an end-to-end empirical evaluation of the deadline-Information Leakage Score (ILS-dl) extension introduced in the companion methodology paper. The deadline-ILS extends the original ILS to deadline-resolved prediction-market contracts, the dominant structural form of publicly documented insider trading on Polymarket. We anchor the evaluation in the 2026 U.S.-Iran conflict cluster of the ForesightFlow Insider Cases (FFIC) inventory, the largest documented deadline cluster. The evaluation has four parts: per-category exponential-hazard estimation, a single-case ILS-dl computation, cross-market wallet analysis, and methodological refinements. Hazard-rate estimation produces an adequate exponential fit for military-geopolitics markets (KS p = 0.609, half-life 2.3 days) and a preliminary fit for corporate-disclosure markets (n = 5). The regulatory-decision category is rejected as bimodal (p = 0.013). On the largest applicable FFIC contract ("US forces enter Iran by April 30," $269M volume), the article-derived public-event timestamp yields ILS-dl = +0.113 versus a resolution-anchored proxy value of -0.331: a 0.444 shift in magnitude on opposite sides of zero, demonstrating that the extension distinguishes signal from proxy artefact. Pre-event drift is mild, and short-window variants (30-min, 2-hour) are exactly zero. Cross-market wallet analysis identifies 332 wallets active in both major Iran-cluster markets, but the available trade history covers only the resolution-settlement window.

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.

Desk Editor's Note private letter to a colleague

The paper shows a sign flip in ILS-dl on the largest Iran contract when switching from resolution proxy to article-derived timestamp, but the contrast rests on an unvalidated event date with no sensitivity check.

read the letter

The main thing to know is that this paper finds a clear sign flip in the deadline-ILS on the largest Iran contract when using article timestamps instead of resolution proxies, but that result is driven by the unverified choice of when the public event happened. The work applies the ILS-dl extension to the documented FFIC cases around U.S.-Iran 2026 markets. It estimates exponential hazard rates by category, shows an adequate fit for military-geopolitics markets with a KS p-value of 0.609 and half-life of 2.3 days, and rejects the regulatory category for bimodality. On the $269M volume contract, it reports ILS-dl of +0.113 with the article time versus -0.331 with the proxy, which is the headline contrast. The wallet overlap check finds 332 shared wallets but notes the trade history is limited to the resolution period. This is useful as a first empirical check on the method extension. The hazard fits and the specific ILS numbers give something concrete to look at, and the short-window variants being zero is a nice detail. The soft spots are mostly around the data and assumptions. The corporate category rests on just five cases, which is too small for much weight. The key sign difference depends entirely on the article-derived timestamp without any sensitivity analysis or cross-check against primary sources. Shifting that date by a day would change the pre-event window and likely the result. The FFIC cases are taken as given without external confirmation they are real insiders. The circularity in fitting the hazard to the data and then using it for ILS is there but probably minor. This paper is for readers already interested in prediction market monitoring tools or information leakage in betting platforms. It won't move the broader field but provides a practical test case. I would send it to peer review. The empirical application is straightforward and the numbers are reported clearly enough that referees can assess the timestamp issue directly. A revision could add robustness checks on the event times.

Referee Report

3 major / 1 minor

Summary. The paper empirically evaluates the deadline-ILS (ILS-dl) extension on the 2026 U.S.-Iran conflict cluster from the ForesightFlow Insider Cases (FFIC) inventory. It performs per-category exponential hazard-rate estimation (with KS tests), computes ILS-dl on the largest applicable contract using an article-derived public-event timestamp, conducts cross-market wallet analysis, and proposes methodological refinements. The central numerical result is ILS-dl = +0.113 versus a resolution-anchored proxy of -0.331 on the $269M 'US forces enter Iran by April 30' contract, presented as evidence that the extension separates leakage signal from proxy artefact.

Significance. If the central numerical contrast holds under robustness checks, the work supplies the first end-to-end empirical test of deadline-resolved ILS on documented Polymarket cases, including category-specific hazard fits with p-values and a concrete demonstration of sign reversal relative to proxy. This would strengthen the case for using ILS-dl in high-volume deadline markets, though the small n for corporate markets and data limitations on wallet histories constrain broader claims.

major comments (3)

[Abstract] Abstract (single-case ILS-dl computation): the headline result of ILS-dl = +0.113 versus proxy -0.331 on the $269M Iran contract is obtained from an article-derived public-event timestamp that defines the pre-event window for the exponential-hazard model. No sensitivity analysis to plausible shifts in this timestamp (e.g., ±1 day) or independent primary-source validation is reported; such a shift would change the hazard-rate input and could eliminate the reported opposite-sign contrast that is offered as evidence the extension distinguishes signal from artefact.
[Abstract] Abstract (hazard-rate estimation): the corporate-disclosure category fit rests on n=5 observations, which is insufficient to support reliable exponential-hazard claims or cross-category comparisons; the regulatory-decision category is rejected for bimodality (p=0.013) but the implications for the overall ILS-dl evaluation are not quantified.
[Abstract] Abstract (cross-market wallet analysis): the reported 332 wallets active in both major Iran-cluster markets have trade history limited to the resolution-settlement window, so no pre-event leakage patterns can be examined; this data limitation undermines any claim that wallet-level evidence corroborates the ILS-dl result.

minor comments (1)

[Abstract] Abstract: no error bars, confidence intervals, or robustness checks are described for the hazard-rate estimates or the ILS-dl value; adding these would improve interpretability of the reported p-values and numerical contrast.

Simulated Author's Rebuttal

3 responses · 0 unresolved

Thank you for the referee's insightful comments on our paper. We provide point-by-point responses below and indicate the revisions we will implement.

read point-by-point responses

Referee: [Abstract] Abstract (single-case ILS-dl computation): the headline result of ILS-dl = +0.113 versus proxy -0.331 on the $269M Iran contract is obtained from an article-derived public-event timestamp that defines the pre-event window for the exponential-hazard model. No sensitivity analysis to plausible shifts in this timestamp (e.g., ±1 day) or independent primary-source validation is reported; such a shift would change the hazard-rate input and could eliminate the reported opposite-sign contrast that is offered as evidence the extension distinguishes signal from artefact.

Authors: The public-event timestamp is derived from contemporaneous news articles reporting the event, which is a standard approach in such empirical studies. However, we agree that sensitivity analysis would strengthen the result. In the revised version, we will include checks shifting the timestamp by ±1 and ±2 days, recompute the exponential hazard rates and ILS-dl, and report whether the 0.444 magnitude shift and sign reversal persist. We will also attempt to cross-validate the timestamp with additional sources. revision: yes
Referee: [Abstract] Abstract (hazard-rate estimation): the corporate-disclosure category fit rests on n=5 observations, which is insufficient to support reliable exponential-hazard claims or cross-category comparisons; the regulatory-decision category is rejected for bimodality (p=0.013) but the implications for the overall ILS-dl evaluation are not quantified.

Authors: We acknowledge the small sample size (n=5) for corporate-disclosure markets and have described it as preliminary in the manuscript. The central ILS-dl result is computed using the military-geopolitics hazard rate (KS p=0.609), which has sufficient observations. We will revise the abstract and discussion to explicitly state that the primary finding does not rely on the corporate fit and to quantify that the bimodality rejection for regulatory-decision (p=0.013) does not affect the reported ILS-dl contrast, as that category is not used in the computation. This addresses the implications for the overall evaluation. revision: yes
Referee: [Abstract] Abstract (cross-market wallet analysis): the reported 332 wallets active in both major Iran-cluster markets have trade history limited to the resolution-settlement window, so no pre-event leakage patterns can be examined; this data limitation undermines any claim that wallet-level evidence corroborates the ILS-dl result.

Authors: The manuscript reports the 332 overlapping wallets but explicitly notes that trade history is limited to the resolution-settlement window and does not claim this provides evidence of pre-event leakage patterns or directly corroborates the ILS-dl result. The wallet analysis serves to document market participation overlap. We will revise the relevant section to clarify this scope and emphasize the data limitation, ensuring no overstatement of its implications. revision: yes

Circularity Check

1 steps flagged

Minor self-citation to companion methodology; central empirical result independent of inputs

specific steps

self citation load bearing [Abstract]
"This paper reports an end-to-end empirical evaluation of the deadline-Information Leakage Score (ILS-dl) extension introduced in the companion methodology paper."

The central evaluation relies on the ILS-dl definition from the companion paper (same author). While the numerical result itself is computed from FFIC data and timestamps rather than being definitionally equivalent, the extension's validity is presupposed via self-citation without independent verification in this manuscript.

full rationale

The paper applies the deadline-ILS extension (defined in a companion methodology paper by the same author) to external FFIC inventory cases and article-derived timestamps. Hazard rates are fitted per category with reported KS tests and half-lives, then used to compute ILS-dl on one contract, producing an empirical numerical contrast (+0.113 vs. -0.331) against a resolution-anchored proxy. This contrast is not forced by construction from the fit or timestamp choice; it is presented as an observed outcome. The self-citation is limited to method definition and is not load-bearing for the empirical claim, which rests on external data and independent proxy comparison. No reduction of the reported ILS-dl sign flip to the fitted parameters or timestamps occurs by the paper's own equations.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The evaluation rests on fitted exponential hazard rates for resolution timing and the assumption that FFIC cases represent genuine insider activity; no new entities are postulated.

free parameters (1)

exponential hazard rate = half-life 2.3 days
Fitted to military-geopolitics market resolution data, yielding half-life of 2.3 days.

axioms (1)

domain assumption Resolution times follow an exponential distribution
Invoked for per-category hazard-rate estimation in the military-geopolitics and corporate-disclosure categories.

pith-pipeline@v0.9.0 · 5588 in / 1305 out tokens · 45019 ms · 2026-05-08T01:30:42.593227+00:00 · methodology

discussion (0)

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Manipulation, Insider Information, and Regulation in Leveraged Event-Linked Markets
q-fin.TR 2026-05 unverdicted novelty 7.0

Leverage scales market-price manipulation linearly while shifting outcome-manipulation thresholds and multiplying informed-trading rents in three distinct ways, calling for re-allocated regulatory attack surfaces rath...
A Taxonomy of Event-Linked Perpetual Futures: Variant Designs Beyond the Single-Market Binary Case
q-fin.TR 2026-05 unverdicted novelty 6.0

The paper organizes seven canonical variants of event-linked perpetual futures along four design axes, supplying payoff definitions, inheritance rules from prior work, and variant-specific constraints.
Fill-Side Non-Retail Trading on Polymarket: An Empirical Study of Behavioral Tiers and Microstructure Signatures Under Quote-Attribution Constraints
q-fin.TR 2026-05 conditional novelty 5.0

Polymarket fill-side trading appears uni-modal due to missing quote-lifecycle data, with whale, high-frequency, and power-trader tiers dominating 81.4% of notional across 12.6% of addresses.

Reference graph

Works this paper leans on

38 extracted references · 1 canonical work pages · cited by 3 Pith papers

[1]

Journal of Economic Perspectives , volume =

Wolfers, Justin and Zitzewitz, Eric , title =. Journal of Economic Perspectives , volume =
[2]

and Forsythe, Robert and Gorham, Michael and Hahn, Robert and Hanson, Robin and Ledyard, John O

Arrow, Kenneth J. and Forsythe, Robert and Gorham, Michael and Hahn, Robert and Hanson, Robin and Ledyard, John O. and Levmore, Saul and Litan, Robert and Milgrom, Paul and Nelson, Forrest D. and Neumann, George R. and Ottaviani, Marco and Schelling, Thomas C. and Shiller, Robert J. and Smith, Vernon L. and Snowberg, Erik and Sunstein, Cass R. and Tetlock...
[3]

Surowiecki, James , title =
[4]

Information Systems Frontiers , volume =

Hanson, Robin , title =. Information Systems Frontiers , volume =
[5]

The Journal of Prediction Markets , volume =

Hanson, Robin , title =. The Journal of Prediction Markets , volume =
[6]

and Servan-Schreiber, Emile and Tetlock, Philip and Ungar, Lyle and Mellers, Barbara , title =

Atanasov, Pavel and Rescober, Phillip and Stone, Eric and Swift, Samuel A. and Servan-Schreiber, Emile and Tetlock, Philip and Ungar, Lyle and Mellers, Barbara , title =. Management Science , volume =
[7]

and Stiglitz, Joseph E

Grossman, Sanford J. and Stiglitz, Joseph E. , title =. American Economic Review , volume =
[8]

Trading volume statistics , year =
[9]

, title =

Kyle, Albert S. , title =. Econometrica , volume =
[10]

and Milgrom, Paul R

Glosten, Lawrence R. and Milgrom, Paul R. , title =. Journal of Financial Economics , volume =
[11]

and O'Hara, Maureen and Paperman, Joseph B

Easley, David and Kiefer, Nicholas M. and O'Hara, Maureen and Paperman, Joseph B. , title =. Journal of Finance , volume =
[12]

and O'Hara, Maureen , title =

Easley, David and Kiefer, Nicholas M. and O'Hara, Maureen , title =. Review of Financial Studies , volume =
[13]

Journal of Finance , volume =

Easley, David and Hvidkjaer, Soeren and O'Hara, Maureen , title =. Journal of Finance , volume =
[14]

Journal of Financial Economics , volume =

Duarte, Jefferson and Young, Lance , title =. Journal of Financial Economics , volume =
[15]

Flow Toxicity and Liquidity in a High-Frequency World , journal =

Easley, David and L. Flow Toxicity and Liquidity in a High-Frequency World , journal =
[16]

and Bondarenko, Oleg , title =

Andersen, Torben G. and Bondarenko, Oleg , title =. Journal of Financial Markets , volume =
[17]

and Ready, Mark J

Lee, Charles M.\ C. and Ready, Mark J. , title =. Journal of Finance , volume =
[18]

, title =

Hawkes, Alan G. , title =. Biometrika , volume =
[19]

Quarterly Journal of Economics , volume =

Budish, Eric and Cramton, Peter and Shim, John , title =. Quarterly Journal of Economics , volume =
[20]

, title =

Platt, John C. , title =. Advances in Large Margin Classifiers , publisher =. 1999 , pages =

1999
[21]

Proceedings of the 22nd International Conference on Machine Learning (ICML) , pages =

Niculescu-Mizil, Alexandru and Caruana, Rich , title =. Proceedings of the 22nd International Conference on Machine Learning (ICML) , pages =
[22]

arXiv preprint arXiv:2508.03474 , year =

Saguillo, Oriol and Ghafouri, Vahid and Kiffer, Lucianna and Suarez-Tangil, Guillermo , title =. arXiv preprint arXiv:2508.03474 , year =

work page arXiv
[23]

2026 , howpublished =

Mitts, Joshua and Ofir, Moran , title =. 2026 , howpublished =

2026
[24]

2020 IEEE Symposium on Security and Privacy (S&P) , pages =

Daian, Philip and Goldfeder, Steven and Kell, Tyler and Li, Yunqi and Zhao, Xueyuan and Bentov, Iddo and Breidenbach, Lorenz and Juels, Ari , title =. 2020 IEEE Symposium on Security and Privacy (S&P) , pages =

2020
[25]

and Tetlock, Philip E

Schoenegger, Philipp and Tuminello, Peter S. and Tetlock, Philip E. and Karger, Ezra and Atanasov, Pavel , title =. Science Advances , volume =. 2024 , doi =

2024
[26]

, title =

Leetaru, Kalev and Schrodt, Philip A. , title =. Annual Convention of the International Studies Association , year =
[27]

2026 , month = apr, note =

Nechepurenko, Maksym , title =. 2026 , month = apr, note =

2026
[28]

2026 , note =

Nechepurenko, Maksym and Shuvalov, Pavel , title =. 2026 , note =

2026
[29]

Soros, George , title =
[30]

, title =

Schelling, Thomas C. , title =
[31]

, title =

Manski, Charles F. , title =. Economics Letters , volume =
[32]

2026 , note =

Tsang, Kwok Ping and Yang, Zichao , title =. 2026 , note =

2026
[33]

and Peng, L

Ng, H. and Peng, L. and Tao, Y. and Zhou, D. , title =. 2026 , note =

2026
[34]

and Huang, Taeyoung , title =

Clinton, Joshua D. and Huang, Taeyoung , title =. 2025 , note =

2025
[35]

, title =

Murphy, Allan H. , title =. Journal of Applied Meteorology , volume =
[36]

, title =

Brier, Glenn W. , title =. Monthly Weather Review , volume =
[37]

, title =

Gneiting, Tilmann and Raftery, Adrian E. , title =. Journal of the American Statistical Association , volume =
[38]

2026 , note =

Nechepurenko, Maksym , title =. 2026 , note =

2026