Stellar RADAR
A surveillance engine purpose-built for event-driven and prediction markets. RADAR scores every trade as it happens and flags behavior that does not match an honest participant, so exchanges, market operators, and compliance teams can act on potential insider trading and manipulation while it still matters.
Prediction Markets Break Traditional Surveillance
Equity market surveillance listens for signals that only appear in dense, historical trade data. Event-driven markets do not work that way.
Built for traditional markets
- Long trading histories to compare against
- A stable set of instruments that trade for years
- Baselines built from months of prior activity
- Alerts reviewed after the fact, often long after settlement
What event-driven markets actually look like
- Thousands of newly created, short-lived contracts
- Little or no trading history on most instruments
- Outcomes that resolve in days or hours
- Information advantages that are tied to real-world events
Detection by Behavioral Reflection, Not Confession
RADAR never needs an insider to admit anything. It scores every trade and lets deviation from ordinary behavior stand out on its own.
Every Trade, As It Happens
RADAR sweeps the market continuously rather than sampling it. Each trade is scored the moment it occurs, so potential misconduct surfaces before the contract settles.
Context From Many Angles
Each score draws on account context, trade context, market activity, and post-trade conduct. No single signal decides the outcome, which keeps false positives low for ordinary participants.
One Transparent Probability
Every trade receives a single probability score, and every flag shows the factors that drove it. Analysts see why a trade was flagged, not just that it was.
Tuned to Each Market
The model is calibrated on the operator's own historical trade data and recalibrated as the market grows, so thresholds reflect how that market actually behaves.
Two Decades of Probabilistic Risk Modeling
RADAR extends a probabilistic surveillance architecture that members of the Stellar leadership team first built for insider trading and fraud detection in U.S. equity markets more than 25 years ago. The same core approach has since been applied across other regulated domains with strong classification performance.
The team pairs that modeling depth with frontline experience as federal investigators, financial regulators, and investment bankers. RADAR is designed by people who have built surveillance systems and worked the cases those systems produce.
Transparent, Auditable, Defensible
A single, interpretable model rather than a black box
Every score traceable to its contributing factors
Built to support referrals and enforcement review
Starts with the data available today and expands as data access grows
Built for the People Responsible for Market Integrity
Exchanges and Operators
Real-time surveillance across every listed contract, including newly created markets with no trading history.
Regulators
An explainable, consistent method for identifying trades that warrant inquiry across event-driven markets.
Compliance Teams
Prioritized alert queues with the rationale attached, so analysts spend time on the trades that matter.
Investigators
Scored, documented trade histories that support case development and referrals.
From First Conversation to Live Pilot
Scope the Data
Identify what trade and account data is available today and what a data-sharing arrangement would add.
Calibrate the Model
Fit and validate RADAR on the operator's historical trade data so scores reflect that market's real behavior.
Run a Bounded Pilot
Score a fixed set of contracts over a fixed window and validate flagged trades against known outcomes.
Go Live
Deploy continuous scoring with analyst queues, reporting, and periodic recalibration as the market grows.