Skip to main content

Signal Engine

Proprietary Signal System​

The Sigma Engine includes a proprietary signal system that monitors over 50+ data endpoints across multiple exchanges and data sources. It produces a directional bias ranging from -1.0 (strong short) to +1.0 (strong long), which is used by all trading strategies to adjust their behavior.

What It Monitors​

The signal engine aggregates data from multiple categories to form a comprehensive market view:

  • Order flow analysis — Real-time monitoring of buying and selling pressure across venues
  • Market microstructure — Orderbook dynamics, depth changes, and structural patterns
  • Cross-exchange data — Divergences between centralized and decentralized venues
  • Volatility regime — Current and predicted volatility conditions
  • Market structure levels — Support/resistance, value areas, liquidation clusters
  • Sentiment indicators — Funding rates, open interest changes, market positioning

Multi-Exchange Data Pipeline​

The engine aggregates real-time feeds from multiple centralized and decentralized venues. Cross-venue analysis surfaces microstructural divergences that typically precede large directional moves, giving the engine a structural information advantage over single-exchange strategies.

Adaptive Weighting​

Signal weights are derived from recursive feature elimination with cross-validation over 360+ days of minute-level tick data, testing nearly 200+ features and fine-tuning down to only those with statistically validated predictive power. Weights are asset-specific because different assets respond to different market dynamics.

Market Regime Detection​

The engine classifies the current market into different regimes and adjusts strategy behavior accordingly. During high-danger regimes, strategies automatically widen spreads and reduce exposure. During favorable conditions, they can tighten spreads and increase activity.