Quant Models
Operations-research and stochastic-process models over the ASX universe:
Ledoit–Wolf shrunk covariance, quadratic-programming optimisation, hierarchical
risk parity, Engle–Granger cointegration and Kelly sizing.
Method notes in docs/QUANT_RESEARCH.md. Research simulation — not financial advice.
Run
docker compose exec web python manage.py run_allocation_backtest --all
to populate this comparison.
Click any frontier point or model-portfolio diamond above to inspect its full weight distribution here (hover shows the top holdings inline).
Grey dots: individual stocks (annualised trailing return vs volatility). Amber line: the long-only efficient frontier. Dashed cyan: the capital allocation line from the risk-free rate through the tangency (max-Sharpe) portfolio. Model portfolios are marked — note how far the naive single stocks sit from the frontier: that gap is the mathematical value of diversification. Trailing returns are noisy estimates, so the frontier moves as the data does.
Run
docker compose exec web python manage.py scan_breakouts
(also runs nightly at 20:00).
Run
docker compose exec web python manage.py optimize_portfolio.
Run
docker compose exec web python manage.py find_pairs.
Run a momentum backtest first
(python manage.py run_backtest).