Skip to content

chore(QC): cleanup redundant projects with poor backtests #141

@jsboige

Description

@jsboige

Context

The QC cloud workspace has 30+ projects, many of which are redundant with each other or have poor backtest results that don't warrant further iteration.

Objective

Reduce project clutter by archiving/removing projects that:

  • Are largely redundant with better-performing versions (e.g., DualMomentum vs DualMomentumNoTLT)
  • Have negative or near-zero Sharpe with no clear path to improvement
  • Were exploratory and the exploration concluded negatively

Candidates for cleanup

Project Sharpe Reason
PairsTrading -0.361 OLS hedge ratio doesn't save it
ForexCarry -0.324 USD trend filter degrades momentum
ETF-Pairs -0.706 Exploratory, failed
VIX-TermStructure 0.051 Near-zero, no alpha
TurnOfMonth 0.128 Calendar anomaly too weak
DualMomentum 0.350 Superseded by DualMomentumNoTLT
TrendFilteredMeanReversion -0.016 Cash drag, ceiling reached

Process

  1. Review each candidate - confirm no unique insight worth preserving
  2. Document lessons learned from each in the patterns file
  3. Archive projects (don't delete - keep for reference)
  4. Close associated iteration issues (QC ForexCarry: Improve robustness (Sharpe -0.654) #17, QC VIX-TermStructure: FIXED (Sharpe +0.05) #18, QC TurnOfMonth: Improve robustness (Sharpe 0.127) #21, etc.)
  5. Update strategy catalog

Related issues

Metadata

Metadata

Assignees

No one assigned

    Labels

    priority-mediumNEEDS_IMPROVEMENT strategiesquantconnectQuantConnect strategy development

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions