TOOLS / PORTFOLIO OPTIMIZER

What is the right mix?

Five solvers over the full asset universe, on forward-looking assumptions rather than a backtest of the last four decades. It shows the frontier, the trade, and every position that moves.

FREE · NO SIGNUP · RUNS IN YOUR BROWSER

Choose a Starting Point

Choose Your Goal

Optimized Portfolio

Expected Return

7.6%

Volatility

8.2%

Sharpe Ratio

0.583

Est. Max Drawdown

-20.4%

Allocation

Asset	Weight	Return	Vol	Constrain
US Aggregate Bonds	64.7%	4.8%	4.8%	-%
US REITs	16.4%	8.8%	17.4%	-%
Bitcoin	11.6%	15.0%	42.5%	-%
Gold	7.3%	5.5%	16.7%	-%

Set min/max constraints on any asset and the portfolio re-optimizes automatically.

Method: Maximum Sharpe. Risk-free rate: 3.1% (US Cash, JPM LTCMA 2026). Cash sat out this search: at the risk-free rate it cannot raise the Sharpe ratio, only scale the risk, so it holds only what a rule or a volatility cap called for.Max drawdown estimated as 2.5× annual volatility.

See where your portfolio sits on the efficient frontier

Visualize the full risk-return tradeoff across all feasible portfolios.

Methodology

Data source: Expected returns, volatilities, and correlations from J.P. Morgan 2026 Long-Term Capital Market Assumptions (30th edition, pages 82-83, USD matrix). Bitcoin assumptions are Portfolio Lab estimates (15% geometric return, 42.5% volatility) informed by institutional research. Read our methodology.

Max Sharpe: Grid search over target returns along the efficient frontier, solving for minimum variance at each target. Selects the portfolio with the highest Sharpe ratio.

Min Variance: Projected gradient descent with box constraints, minimizing total portfolio variance (w'Σw).

Risk Parity: Equal Risk Contribution (ERC) via cyclical coordinate descent (Griveau-Billion et al., 2013). Each asset contributes equally to total portfolio risk.

HRP: Hierarchical Risk Parity (López de Prado, 2016). Single-linkage clustering on correlation distance, then recursive bisection weighted by inverse variance.

Black-Litterman: Implied equilibrium returns (π = δΣw) blended with investor views using Bayesian updating.

Constraints: Per-asset minimum and maximum weights. All weights must sum to 100%. No leverage or short-selling.

This tool uses forward-looking assumptions, not historical data. Past performance does not guarantee future results. Not financial advice. Consult a qualified adviser before making investment decisions.

Questions this page answers

What is portfolio optimization?

Portfolio optimization (portfolio optimisation in British English) finds the best mix of assets to maximize return for a given level of risk, or minimize risk for a given return target. Modern portfolio theory (Markowitz, 1952) shows that diversification across imperfectly correlated assets improves risk-adjusted returns. This tool optimizes across up to 57 asset classes using the USD reference assumptions. The signed-in workstation lets you compare supported forecasting sources.

Which portfolio optimization method should I use?

It depends on your goal. 'Best risk-adjusted return' (Max Sharpe) finds the portfolio with the highest return per unit of risk, which suits most investors. 'Smoothest ride' (Min Variance) minimizes volatility, which suits conservative investors and retirees. 'Balanced risk' (Risk Parity) ensures no single asset dominates your risk, which suits all-weather portfolios. 'Data-driven clusters' (HRP) groups correlated assets and weights by stability, which helps when you distrust return forecasts. 'Blend your views' (Black-Litterman) lets you adjust market expectations, which suits investors with strong convictions.

How much Bitcoin should I put in my portfolio?

The modelled allocation depends on expected returns, risk estimates, your objective and the limits you choose. It is not a universal recommended percentage. Bitcoin inputs are Portfolio Lab assumptions. Compare the result with your cash needs and a stress test, and try alternative return assumptions in the workstation.

What data does this optimizer use?

Expected returns and volatilities are J.P. Morgan's 2026 dollar assumptions, marked to today's prices; correlations are measured from the assets' own monthly returns. The full app runs in your base currency on the average of J.P. Morgan, BlackRock, Research Affiliates and AQR, with any single house a switch away.

Keep reading

The free optimizer that uses institutional data →

RESEARCH · THE SAME INPUTS, WITHOUT THE TERMINAL

How much Bitcoin should be in your portfolio? →

RESEARCH · WHAT THE OPTIMIZER SAYS, AND WHY

Portfolio Lab vs Portfolio Visualizer →

COMPARISON · WHAT EACH ONE DOES BETTER

Full methodology →

METHODOLOGY · EVERY FORMULA AND SOURCE

This page solves one portfolio. The workstation saves them, simulates them, and reports on them.

Open the workstation →