Joshua Della Vedova

DV-PMI Surveillance · In development

Matched / Pre-Arranged Orders

Detection of coordinated bilateral trading via network-cluster identification of counterparties who trade primarily with each other. A structural counterparty-persistence screen was previously surfaced here; we have moved it back to development pending implementation of an approximately-closed-cluster method consistent with Sirolly, Ma, Kanoria, and Sethi (Columbia Business School, 2025). The raw structural-screen output remains available for download below.

In development. The structural counterparty-persistence screen (Tier 3 v1) is necessary but not sufficient for wash-cluster identification and was dominated by competing market makers in practice. A network-cluster implementation consistent with Sirolly et al (2025) is the right next step.

Status: in development.

What this index will test

Wallet pairs and small clusters that have been counterparties on many matches across many markets and rarely trade with anyone outside the cluster. The "approximately closed" structure is the signature of coordinated bilateral trading and is what distinguishes it from market-maker rotation, where MMs face many distinct counterparties and concentrate their activity on no single wallet.

Reference implementation

Sirolly, Ma, Kanoria, and Sethi (Columbia Business School, 2025) propose an iterative network-clustering procedure for wash-trade detection and apply it to Polymarket. They estimate that transaction patterns indicative of wash trading peaked at approximately 60 percent of weekly volume in December 2024, persisted through April 2025, and reached approximately 20 percent again in October 2025. Their method is the standard against which our forthcoming implementation will be benchmarked.

Why the previous version was demoted

The earlier release on this page surfaced a structural counterparty-persistence screen with three thresholds (loose, medium, strict). Inspection of the top flagged pairs showed they were dominated by competing algorithmic market makers operating in the same niche markets, not coordinated wash clusters. The structural screen is a necessary precondition for cluster-wash detection (no coordinated activity without persistent counterparty relationships), but it is not sufficient on its own. Surfacing it as a live wash-trading number understated the published evidence on Polymarket and was inconsistent with the site's "patterns consistent with" posture rule. We have moved it to development pending the network-cluster refinement.

Data (structural-screen output, for reproducibility)

The previous structural-screen output remains downloadable for anyone who has cited it. It is not a wash-trade flag set.

surveillance_matched_latest.json · surveillance_matched_top_pairs.csv (top 20 persistent pairs; subject to the methodology page's usage notice)

Other surveillance indices

← All surveillance indices · Informed Trading (PII) · Insider Timing · Adverse Selection · Resolution Surprise · Wash Trading (Tier 1) · Concentration

Cite this index

@misc{dellavedova2026matched,
  title        = {Matched / Pre-Arranged Orders Index (in development)},
  author       = {Della Vedova, Joshua},
  year         = {2026},
  url          = {https://jdellavedova.com/surveillance/matched-orders}
}