Sigil Strategies at scale.
A high-speed data processing engine that iterates through hundreds of trading strategies against enriched market data. Born from Arcana to run independently — decoupled for stability, optimized for speed.
Sigil is live and currently iterating on data processing efficiency and speed, processing hundreds of trading strategies against enriched market data from Arcana.
What it does
The strategy execution layer — where enriched data meets trading logic at scale.
Strategy Processing
Iterates through hundreds of trading strategies against enriched market data at high speed.
High-Speed Iteration
Optimized data processing pipeline designed for rapid strategy backtesting and signal generation.
Born from Arcana
Separated from Arcana for independent stability. Reads Arcana's output but runs on its own lifecycle.
Configurable Pipeline
Strategy definitions are modular and configurable. Add, remove, or modify strategies without touching the core engine.
Performance-Optimized
Batch processing, memory-efficient data access patterns, and profiling-driven optimizations for maximum throughput.
Private Software
Sigil is proprietary due to the sensitive nature of trading strategy implementations. Updates published via blog.
The story
Sigil started as a module inside Arcana. As the project grew, it became clear that data ingestion and strategy processing have fundamentally different concerns — different failure modes, different performance profiles, different release cycles.
Separating Sigil from Arcana was the kind of decision that feels obvious in hindsight. If Sigil crashes mid-strategy, Arcana keeps ingesting. If Arcana restarts during a backfill, Sigil can process from persisted data. Independent scaling, cleaner testing, focused optimization.
Now Sigil consumes Arcana's enriched bar data — tick bars, volume bars, dollar bars, imbalance bars — and runs it through configurable strategy definitions to produce trading signals. The current focus is on processing speed: iterating through hundreds of strategies against years of market data fast enough to make rapid experimentation practical.
Sigil is private software. The trading strategy implementations are sensitive, and I don't want to be responsible for anyone building their own algorithm from this codebase. Progress and case studies are shared on the blog.
Language
Python
Data Source
Arcana enriched bar data (tick, volume, dollar, imbalance bars)
Focus Areas
Strategy processing, data enrichment, signal generation, processing speed
Status
Private — updates via blog posts