Measure the portfolio you have.
Or let the engine search for the best-scoring one.
Two tools, one complete report: return, risk, risk-adjusted ratios and diversification for every portfolio.
2,000+ assets · Monthly prices, from 2000 where available · Free today, no sign-up
Get a full report on your portfolio
Type in your assets and their weights — the portfolio you already hold or one you are considering — and get its complete report in seconds.
Find the best-scoring portfolios
Pick the assets and an objective. The engine explores a very large number of equal-weight combinations, size by size, keeps the best-scoring ones and checks them on a later period kept out of the search.
What you get for every portfolio
Return & risk
How much it grew and how bumpy it was.
Risk-adjusted
Return for each unit of risk taken.
Versus the market
Compared with a benchmark index.
Tail risk
What a bad month can look like.
Diversification
Is the risk really spread out?
Allocation
Sector, region, asset class.
Search on the past, test on a later period
Maximum Sharpe ratio, minimum variance, target volatility, risk parity, minimum CVaR or maximum diversification. There is also a fixed momentum rule, which selects assets instead of optimising.
Asset types (stocks, ETFs, bond ETFs, commodities, funds), number of assets, maximum correlation between assets and sector cap. Optionally, a risk limit relative to an index's volatility, with an editable margin.
It builds a greedy selection (adding, one at a time, the asset that improves the objective most) and then runs a genetic search (it combines and varies the best portfolios over many rounds, keeping the ones that score higher) for each portfolio size, with equal weights. Quick takes about 1 minute, Standard about 5 and Deep up to 30; it stops on its own when it stops improving, and you can press Stop and keep what it found.
Each portfolio is shown with its in-sample and out-of-sample figures side by side, next to two references (the index and the equal-weight universe), a noise reference (what a short version of the same search already reaches on simulated data with no signal) and a stability count (how many repeated random starts ended on the same set of assets).
Analyse or optimise
| 1 · Analyse my portfolio | 2 · Explore with the engine | |
|---|---|---|
| You provide | Assets and their weights | A list of candidate assets, an objective and your limits |
| You get | The full report of that exact portfolio | The best-scoring portfolios and, for each number of assets, the one with the highest in-sample score, each with its full report, and how they held up out-of-sample next to the index |
| Periods | The period you choose | In-sample (search) and out-of-sample (test), side by side |
| Use it to | Understand the risk of what you hold | Explore which combinations scored best on past data and how they held up out-of-sample |
Before you start
Which assets can I use?
Over 2,000 assets: stocks, ETFs, bond ETFs, funds and commodities. Stocks from the US (S&P 500, Nasdaq-100), Europe (Euro Stoxx 50, IBEX 35, DAX, CAC 40, FTSE 100, SMI, AEX and most of the FTSE MIB), Canada (TSX 60), Australia (ASX 50) and Japan (most of the Nikkei 225, listed in Germany, in euros), plus popular ETFs, UCITS ETFs and index funds. Monthly prices, from 2000 where available. Search by name or ISIN.
What exactly does the engine do?
It searches among more than 2,000 stocks, ETFs, bond ETFs, commodities and funds for the equal-weight combinations that score highest on the objective you set, within the limits you set, and then shows how they behaved on a later period that was not used to choose them. Besides the best-scoring portfolios, it shows, for each number of assets, the one with the highest in-sample score, two references (the index and the whole universe at equal weights) and a noise reference that shows scores that chance alone can reach on data with no signal.
Things to know: all weights are equal; data are monthly and price-only (no dividends, costs or taxes); only assets still listed today are included, so past results look better than they were; and the in-sample result is optimistic by construction. The optional risk limit is relative to an index's volatility and is met in the optimisation period; afterwards the risk can be higher or lower, it does not protect against market falls and nothing guarantees it will be met afterwards. The engine does not forecast, does not recommend what to buy and is not investment advice. It does not promise to do better than an index: in our tests, maximising the past Sharpe ratio did not beat the index reliably on later data. Past performance is not a reliable indicator of future results, and you can lose money.
Can I start with my portfolio and then use the engine?
Yes. Build and analyse your portfolio first, then open the engine with the same filters and compare your report with the combinations it found.
Why do in-sample and out-of-sample results differ?
Choices made on past data always look better on that same data. The out-of-sample period shows how the portfolio behaved on a later period kept out of the search, a fairer check. It is still past data, not a forecast.
Plain-language glossary
Greedy selection: the engine adds assets one at a time, each time choosing the one that improves the objective most.
Genetic search: it keeps a population of portfolios, combines and slightly changes the best ones over many rounds, and keeps those that score higher.
HHI (Herfindahl index): the sum of the squared weights, from 0 to 1. The higher it is, the more concentrated the portfolio; 1 means a single asset.
Ulcer index: a measure that combines how deep and how long the falls from previous highs were. A lower value means shallower or shorter falls.
Noise reference: what a short version of the same search already reaches on simulated data with no signal. It helps judge whether a result stands out from chance.
Is it free?
Today you can use the Investing tools free of charge and without creating an account. This may change in the future.
Educational tools and content. Examples use made-up, illustrative numbers. Results from the tools are hypothetical and based on past prices; they are not a forecast, advice or a personal recommendation. Past performance is not a reliable indicator of future results. Investing involves risk of loss.