How to use it 01 Paste one asset per line as Name: r1, r2, … — returns in percent, or prices if you switch the input kind. 02 Choose the frequency (monthly is the default), a risk-free rate and whether short positions are allowed. 03 Read the frontier chart, then the closed-form table: min-variance and tangency returns, volatilities, Sharpe ratios and weights. 04 Type a target return in percent to place that exact point on the frontier, and compare the long-only rows when shorts are forbidden.
Worked readouts The demo is three synthetic monthly series — Equity, Bonds, Gold — over 24 months, with a 2% risk-free rate. Annualised arithmetic means read 12.00% / 4.95% / 9.50% and volatilities 13.08% / 2.31% / 7.48% .
Portfolio
Return
σ
Sharpe
Weights
Minimum variance
5.86%
0.93%
4.147
13.7% / 87.5% / −1.3%
Tangency
6.11%
0.96%
4.28
15.0% / 82.7% / 2.3%
Long-only min variance
5.93%
0.93%
4.21
13.9% / 86.1% / 0%
4,000-sample cloud, best
6.11%
0.96%
4.276
15.1% / 82.9% / 1.9%
The closed-form block explains the curve: A = 11540.1972 , B = 676.3206 , C = 40.7564 , D = 12926.9171 , and the global minimum variance has σ² = 1/A = 0.000087 .
The algebra in one paragraph With expected returns μ and covariance Σ, the frontier minimises wᵀΣw subject to 1ᵀw = 1 and, for a target point, μᵀw = target. The solutions are built from three scalars — A = 1ᵀΣ⁻¹1, B = 1ᵀΣ⁻¹μ, C = μᵀΣ⁻¹μ — with the minimum-variance return at B/A. Adding the risk-free rate moves the best portfolio to the tangency point where a line from the risk-free asset touches the curve (the capital market line); the panel draws that line and reports the tangency as the maximum-Sharpe portfolio. When shorts are forbidden, the closed-form solution is projected onto the long-only simplex and refined iteratively, which is why the long-only rows are separate.
Limits and honest assumptions
Inputs are estimates. Means and covariance come from your sample; mean-variance optimisation is famously sensitive to expected returns, and the engine carries that warning into the card for exactly that reason. Arithmetic annualisation. Monthly means are multiplied by 12 and covariances by 12; compounding uses Metrics , and a 12% arithmetic mean is not a 12% compound return.
The demo is synthetic. Its high Sharpe ratios come from a low-volatility generated series, not from market history.
No costs, taxes or constraints beyond weights summing to 1 and the long-only switch. What it does not do. It does not forecast returns, shrink noisy estimates, or include liabilities; Rebalancing handles drift and costs after you choose weights. This page provides general information and scenario estimates based only on the inputs you provide. It is not investment, tax, legal, or other professional advice and does not recommend any rate, product, or institution.
The minimum-variance solution can short Look at the unconstrained minimum-variance weights: 13.7% Equity, 87.5% Bonds and −1.3% Gold . Nothing is wrong — the maths minimises variance and a small short improves it — but the portfolio is not implementable as displayed, and the long-only row (13.9% / 86.1% / 0%) is the honest alternative. The second trap is reading the demo's Sharpe of 4.28 as an expectation: it is a property of 24 generated months. Estimation error cuts both ways, and a frontier built on 24 observations is a sensitivity test, not a target.
Where it fits Choosing long-run weights Paste several years of monthly returns, compare the tangency and long-only rows, and use the weights as a starting allocation — then stress them: drop the best month, halve the sample, and see whether the ranking survives.
Testing whether a fee earns its keep Solve the frontier with and without a costly asset by adjusting its return series, and read whether the tangency Sharpe improves enough to matter.
Privacy The return series, weights and frontier are computed in your browser; nothing is uploaded or stored.
References
Markowitz, H., Portfolio Selection , The Journal of Finance 7(1), 77–91, 1952, doi.org (访问日期:2026-10-07)— the mean-variance framework.
Wikipedia, Modern portfolio theory , en.wikipedia.org (访问日期:2026-10-07)— diversification, efficient portfolios and estimation error.
Wikipedia, Efficient frontier , en.wikipedia.org (访问日期:2026-10-07)— the frontier curve, tangency and capital market line.
Wikipedia, Sharpe ratio , en.wikipedia.org (访问日期:2026-10-07)— excess return per unit of volatility.
Calculators in this hub
Hand-picked tools, one click away. The mini versions compute live and carry your values into the full calculator.
Sources & review
Reviewed by CalcX Editorial Team
Updated 2026-10-07