What I found
Past a point, more risk stops paying. The efficient frontier draws exactly where that point is: for any level of risk, there is a best achievable return, and chasing return beyond the sensible range just buys volatility.
What I built
A mean-variance model of a $100M equity portfolio across five stocks. It builds the covariance from each asset's volatility and the correlation matrix, traces the frontier, and reports the risk-adjusted ratios at whatever point you choose.
How it works
Drag the risk-tolerance point along the frontier. The allocation and the four ratios, Sharpe, Treynor, Jensen's Alpha, and the Information Ratio, update live.
Allocation of $100M
- Large-cap tech A29%
- Large-cap tech B11%
- Healthcare36%
- Financials17%
- Energy6%
Illustrative sample data, not live prices or investment advice. Risk-free rate 4%; beta and the information ratio use an equal-weight benchmark. Weights are unconstrained, so higher-return points can take a small short position.
The minimum-variance portfolio here sits below the volatility of every single stock in it, which is the whole point of diversification: the mix is steadier than any of its parts.
Impact
- The frontier turns an abstract risk conversation into a concrete choice between named portfolios.
- The ratios say how well each portfolio is paid for the risk it carries, not just how much it might return.
The transferable lesson
The frontier is a decision tool, not a forecast. It narrows the field to the portfolios worth considering at a given risk, and the ratios rank how efficiently each one uses that risk. Judgment still picks the point; the model makes the trade-off honest.