Signals & Models

Transparency in risk detection: types of signals and models used

Liquidity Volatility

Active

Monitors liquidity pool changes, withdrawal rates, and pool balance fluctuations to detect abnormal patterns.

Confidence Level

87%

Wallet Clustering

Active

Identifies coordinated wallet activity, similar transaction patterns, and wallet clusters that may indicate coordinated exits.

Confidence Level

82%

Transaction Timing Anomalies

Active

Analyzes transaction timing patterns, frequency anomalies, and temporal clustering to detect suspicious scheduling.

Confidence Level

75%

LP Ownership Concentration

Active

Tracks liquidity provider ownership distribution to identify concentration risks and potential single-point-of-failure scenarios.

Confidence Level

79%

Contract Change Detection

Experimental

Monitors smart contract upgrades, permission changes, and administrative actions that may indicate potential risks.

Confidence Level

58%

Historical Pattern Matching

Experimental

Compares current behavior patterns against historical rug pull incidents to identify similar pre-exploit signatures.

Confidence Level

65%

Note: Signal confidence levels are based on historical performance and model accuracy. Proprietary detection logic and model weights are not disclosed for security reasons.