Quant Phase 1: Data + Signals + Eval

t-709·WorkTask·
·
·
Created1 month ago·Updated1 week ago·pipeline runs →

Description

Edit

Quant System Phase 1: Data + Signals + Eval

Build the data layer, deterministic signals, and walk-forward eval framework for the speculative fund quant system. No portfolio management, no LLM agents, no web UI — just data access, signal computation, and rigorous evaluation.

Success criteria:

  • Historical daily prices cached locally for 16-instrument universe (2020-present)
  • Momentum signal achieves IC > 0.02 on held-out 2024-2025 data, t-stat > 2.0
  • All signal functions pass null-input and QuickCheck property tests
  • Eval harness runs walk-forward IC measurement and prints pass/fail report

Design doc: Omni/Fund/Quant/DESIGN.md Research docs: ~/ava/ben/research/ai-quant-pipeline-v2.md, signal-agent-evals-v2.md

Timeline (7)

🔄[human]Open → InProgress1 month ago
🔄[human]InProgress → Review1 month ago
💬[human]1 month ago

Phase 1 complete. All 7 subtasks done. First passing signal on real data: 252d momentum IC=0.087, t=2.23. Full parameter scan infrastructure in place. Ready for Phase 2: alpha combination, portfolio optimization, web dashboard.

💬[human]1 week ago

Ava verified: implementation present in live tree (file/feature confirmed via inspection). Moving to Verified.

🔄[human]Review → Verified1 week ago