Structure discovery. Fraud detection.

Market Intelligencefor Financial Integrity

Inspired by
the unknown

GammaRay is a structure-discovery AI for complex networks. In markets, it reconstructs relationships among accounts, trades, venues, and time windows to surface coordinated behavior that isolated alerts can miss.

The same engine has also been successfully applied to dark-matter halo detection in astrophysics.

Inside the technology

Available now

Wash-Trading Detection and Mapping

Identify coordinated account systems behind synthetic volume. Map the trades, counterparties, timing, and relationships that connect them.

Public index (in development)

Wash-Trading Volatility Index

Estimate how much observed price volatility is associated with wash-trading activity in a defined market and period.

Look broaderLook closerLook deeper

GammaRay examines trading activity across markets and time, maps individual account groups, and connects recurring accounts across cases. The analysis presents the findings with transaction records, account roles, and market context for review.

Illustrative networkMarket network

Illustrative case view

Account roles: Mapped
Trade direction: Linked
Timing window: Defined
Group context: In view

A wider field connects accounts across markets and time.

Our technology

Research and
evidence

Our published research presents market findings and measured detection performance. Each evaluation includes its dataset, observation period, and test conditions.

Public and research released in full. Restricted material is shared separately under NDA.

All publications

First used to find dark-matter halos

GammaRay began as a way to reveal structures that only appear when observations are considered together. The same engine has been applied to dark-matter halo detection in astrophysics, and it reads a market the same way: accounts, transactions, and time as one connected system.

The scientific foundation

A method built for hidden structure

Rather than scoring trades one at a time, the engine reconstructs the relationships between them. Each result is reported with its evidence conditions attached: the sample, the observation period, the records available, and the limits of what the finding can support.

Inside the technology

Research that keeps going

Every new setting needs its own data and evaluation. Current work looks at how findings change as the available evidence changes, at transfer across market structures, and at what detected structures can inform next. Detailed methods and validation material are available under NDA.

Ongoing research