portfolio-optimize

portfolio-optimize

Tier: 1 (research, no live money)
Trigger: Monthly, on-demand for subscriber portfolio recommendations
MCP Tool: run_algoclaw_skill("portfolio-optimize", {"bots":["MNQ","CL","MES","NQ"],"capital":50000})
Source Pattern: Riskfolio-Lib HRP + PyPortfolioOpt

What It Does

Computes optimal capital allocation across AlgoChains bots using:

1. HRP (Hierarchical Risk Parity) — groups correlated bots, allocates inversely to risk
2. Min-Variance — minimizes portfolio volatility for given return target
3. Conservative dual — takes minimum weight per bot across both methods

Input

• Bot performance history (pulled from live logs or Supabase)
• Correlation matrix between bot returns
• Subscriber's risk tolerance and capital

Algorithm (pure Python, no Riskfolio dependency required)

# Pure HRP implementation using scipy + numpy
# 1. Compute correlation matrix from bot return histories
# 2. Build hierarchical linkage tree (Ward / single)
# 3. Recursive bisection: allocate capital proportional to cluster variance
# 4. Apply min 5% / max 40% per-bot constraints

Output

{
  "capital": 50000,
  "method": "HRP",
  "allocations": {
    "MNQ": {"weight": 0.38, "capital_usd": 19000, "sharpe": 4.61},
    "CL":  {"weight": 0.28, "capital_usd": 14000, "sharpe": 2.8},
    "MES": {"weight": 0.18, "capital_usd": 9000,  "sharpe": 2.1},
    "NQ":  {"weight": 0.16, "capital_usd": 8000,  "sharpe": 2.3}
  },
  "portfolio_sharpe_est": 3.2,
  "correlation_note": "MNQ+NQ 96% correlated — HRP auto-reduces both",
  "rebalance_frequency": "monthly"
}