MFL-PORTFOLIO-SCIENCE · Quantitative Portfolio Science

How to use Quant 2.0

Five steps to your first result. No background in portfolio theory needed — just follow the clicks below.

Load some price history

In the left sidebar, under "Price history," pick Live fetch, then click Try live fetch (~2y daily).

In a hurry, or the internet's being difficult? Click Demo data → Generate demo history instead — instant, synthetic prices so you can try everything below without waiting on real data.

(Optional) Check your shares

Five ASX shares are already loaded for you at the top of the sidebar. Leave them as-is for your first run, or click + Add a share to build your own basket — up to 12.

Turn on Regime-aware (HMM)

Scroll to "Regime & risk model" → under "Estimation mode," click Regime-aware (HMM).

What this does: instead of treating the market as one unchanging condition, this fits a model that recognizes "calm" vs "stressed" periods in the data, and leans the numbers toward whichever one looks more likely next.

Turn on CVaR

Just below that, under "Risk measure," click CVaR.

What this does: most tools only look at how much a portfolio swings on average. CVaR instead asks "on the worst days, how bad does it actually get?" — a more honest read of real downside risk.

Run it

Click the big Run optimization button at the very bottom of the sidebar.

Reading your result

The Results tab shows two portfolios side by side, built from the exact same data:

Classical MVO The naive textbook approach — trusts the historical average return exactly as-is, no matter how little data backs it up.

Robust SOCP Quant 2.0's own method — assumes your return estimates could be wrong, and picks weights that hold up even in the worst case within that uncertainty.

Curious how well "Robust" actually holds up? Open the Reliability Guide tab at the top of the results — it shows real historical validation, not just this run's numbers.