Acubic vs Composer: Builder or Strategy Engine?

Key takeaways

• Composer runs rules that the investor writes; Acubic builds a portfolio from the investor's risk profile using a published method.
• Composer suits someone with a market thesis to test and automate; Acubic suits someone who wants a diversified, risk-calibrated portfolio without forming a view on timing.
• A good backtest is evidence about the past, not a forecast: strategies chosen as the best of many tested variants often disappoint on new data.
• Acubic keeps your positions at eToro or Trading 212; on a standard eToro account each rebalance needs your one-click approval.
• Neither tool can guarantee an investment outcome, and nothing here is financial advice.

Investors searching for systematic portfolio tools often encounter both Acubic and Composer. Both automate the investment process. Both use data and rules. Both appeal to investors who want something more disciplined than picking stocks by hand. But the job each one does is structurally different, and treating them as interchangeable leads to connecting the wrong tool to the wrong problem.

The difference comes down to where each tool starts. Composer starts with the investor's logic and executes it. Acubic starts with the investor's risk profile and derives the allocation from it. That single distinction shapes everything about how the two tools work.

What Composer is

Composer is a no-code systematic trading platform. It lets investors build conditional strategies using rules they define themselves. Those rules can be based on price signals, momentum, relative performance, or other market conditions. When the conditions are met, the strategy executes automatically, rotating into the allocations the rules specify. Composer refers to these strategies as symphonies.

The design puts the investor's market thesis at the centre. The investor decides what signals matter, which conditions trigger a rotation, and how the allocation changes in each state. Composer's role is to make those rules executable without programming knowledge, and to provide a backtesting environment so the investor can see how the strategy would have behaved historically before running it with real capital.

This makes Composer well suited to investors who have a view about market dynamics and want to express it systematically. The view might be simple: rotate from equities to bonds when a momentum indicator falls below a threshold. Or more complex: a multi-signal model that weights several conditions. Either way, the investor is making a claim that certain conditions in the past predicted something worth acting on, and building a strategy that acts on that claim going forward.

Backtesting does not turn a claim about the future into a guarantee, and no serious use of these tools treats it as one. Past performance of a strategy is evidence about the past. What backtesting provides is a structured way to assess whether a rule held over a historical period and what its drawdown behaviour looked like. That is a useful discipline before running a strategy live.

The main hazard is overfitting: tuning a rule so closely to past data that it captures noise rather than a lasting pattern. A 2014 paper by David Bailey, Marcos López de Prado and colleagues, Statistical Overfitting and Backtest Performance, shows how easily this happens when a computer tests many variants of a strategy and keeps the best one. Such an "optimal" strategy, the authors write, "often performs very poorly 'out of sample' (i.e., on another dataset)". The practical lesson for any rule-based tool is to test few variants, prefer simple rules, and treat a strong backtest as a hypothesis rather than a result.

What Acubic is

Acubic takes a different approach. It does not ask for rules. It asks for a risk profile: how much portfolio concentration you are comfortable holding, your investment time horizon, and how much drawdown you can absorb. From those inputs, the construction method derives an allocation using a published quantitative process.

The method is documented on the Acubic methodology page: a screening and ranking stage selects assets from the investable universe using quality and momentum criteria, an optimiser assigns weights subject to hard constraints derived from the investor's risk inputs, and the result is a specific, inspectable portfolio proposal. The allocation step offers optimisers (hierarchical risk parity, mean-variance and CVaR minimisation) alongside simpler rule weightings such as equal weighting. Nothing is preselected: you choose the method, and the strategy menus for beginner and intermediate investors recommend a rule weighting for each risk level. The inputs translate to explicit constraints on the optimiser, not soft preferences it weighs loosely. A lower risk tolerance produces tighter concentration limits and narrower weight bounds, and those constraints are enforced rather than suggested.

The investor contributes no market thesis. The system builds the portfolio from the risk inputs, not from a view about when to hold or rotate. Once connected to a broker, Acubic monitors the live portfolio for drift and rebalances on the schedule the investor sets. Each rebalance reruns the full construction process from current market data rather than returning the portfolio to a fixed historical basket. The full process, from the first risk inputs through to a live connected portfolio, is covered in the how Acubic works guide.

On a standard eToro connection, each rebalance requires one-click approval: Acubic prepares the orders and the investor confirms before they are submitted. On a Trading 212 connection, and on an eToro sub-account, full automation is available and the scheduled process runs unattended. Those are structural differences between account types, not settings an investor switches on.

The core distinction: conditional logic versus risk-calibrated construction

The clearest way to describe the difference is to look at what each tool asks the investor to provide.

Composer asks for logic. If you can state your investment thesis as a set of conditions and outcomes, Composer can execute that thesis systematically and consistently. The quality of the result depends on the quality of the logic. A well-constructed conditional strategy, properly backtested, is a serious instrument. A poorly constructed one will execute its mistakes reliably and without complaint.

Acubic asks for a risk profile. If you can state how much risk you want to hold and over what horizon, Acubic derives a portfolio that satisfies those constraints using a documented method. The investor does not need a view on when to buy or sell, because the construction process does not require one. It is building a risk-efficient allocation from the investor's inputs, not expressing anyone's view about which assets will outperform.

Composer and Acubic side by side

Question	Composer	Acubic
What you provide	Your own rules and signals	Your risk profile: concentration comfort, horizon, drawdown tolerance
How the allocation is set	By the conditions in your strategy	By a published screening, ranking and allocation method you choose
Conditional trading rules (if/then logic)	Yes, a core feature, with backtesting of those rules	No. Acubic backtests the portfolio it builds, using the filters, ranking order and allocation method you choose
Market view required	Yes, the strategy is your view	No; the method does not forecast short-term moves
When trades happen	When the conditions in your rules are met	On the schedule you set; on a standard eToro account, after your one-click approval

Acubic does not predict short-term market moves. The methodology page states this explicitly. The construction method uses historical covariance to build a portfolio that is diversified and risk-calibrated at the time of construction. What markets do next is not something the model knows or claims to know. Why Acubic does not predict the market covers this in more detail, and it is worth reading before connecting any automated system to a live account.

The reasoning for offering hierarchical risk parity among the optimisers is covered in why Acubic uses hierarchical risk parity. The short version is that HRP handles estimation error better than classic mean-variance when applied to retail-scale portfolios, where the covariance inputs are noisier and the historical window is shorter.

Estimation error is the gap between the statistics an optimiser is fed (expected returns, volatilities and correlations estimated from history) and their true, unknown values. Its size is easy to underestimate. In a 2009 study in the Review of Financial Studies, Victor DeMiguel, Lorenzo Garlappi and Raman Uppal tested 14 optimisation models across seven empirical datasets and found that none was consistently better than simply splitting money equally across the assets. They estimated that a sample-based mean-variance portfolio of 25 assets would need an estimation window of around 3000 months of data before it reliably beat that naive benchmark (Optimal Versus Naive Diversification). That is the problem methods such as HRP are designed to soften, not eliminate.

What each tool is not

Acubic is not a strategy engine. It does not let investors define conditional rules about when to buy or sell, so there is no if/then logic to write or backtest. What Acubic does backtest is the portfolio it builds from the filters, ranking order and allocation method you choose. If what you want from a tool is the ability to build, test, and run your own systematic logic, Acubic does not provide that and is not designed to.

Composer is not a portfolio builder in the sense that Acubic is. It does not take a risk profile as input and derive a specific risk-calibrated allocation from a published quantitative method. It does not monitor portfolio drift against a target produced by that method. An investor who does not have a strategy to express does not have a starting point in Composer. The strategy is what the tool executes, and the investor supplies it.

Neither tool is a robo-advisor in the traditional sense. Robo-advisors typically hold assets in custodial accounts they manage under a regulatory framework for discretionary portfolio management. Acubic connects to the investor's existing eToro or Trading 212 account. Your positions stay at your broker and never move to a new custodian. The structural differences between these models are covered in the AI portfolio builder vs robo-advisor comparison.

Who each tool is designed for

Composer is well suited to an investor who:

• Has a developed view about market conditions and wants to express it as a systematic, backtestable set of rules
• Wants conditional capability: the ability to rotate between asset classes or strategies based on defined signals
• Is comfortable designing and owning the logic, and accepts that the strategy's performance depends on the quality of that logic

Acubic is well suited to an investor who:

• Wants a diversified portfolio built from a documented quantitative method, without needing to form a view on market timing
• Already has an eToro or Trading 212 account and wants to maintain positions there
• Wants the portfolio maintained automatically on a schedule, with an approval step or full automation available depending on their account type and broker
• Wants to inspect the construction method before connecting anything, and verify that it stays within stated constraints

These are often different investors. The investor who finds backtesting a natural part of their process is usually someone who has a hypothesis they want to test systematically. The investor who wants a risk profile turned into a portfolio and maintained over time is usually someone who wants a repeatable process they did not have to design themselves.

Using both

The two tools are not technically exclusive. An investor could run a systematic Composer strategy on one account while using Acubic to manage a separately allocated, risk-calibrated portfolio at their eToro or Trading 212 account. Whether that makes sense depends on keeping the boundaries clear and not adding cognitive complexity that the automation was meant to reduce. For most investors, one coherent approach is more useful than two overlapping ones.

How to decide

The practical question is straightforward: which one asks you to contribute what you actually have? If you are weighing rule-based automation more broadly, the comparison of signal-driven trading bots and portfolio builders covers the same split from a different angle.

If you have a thesis, a set of conditions and rules you believe are worth systematising and testing, Composer is built for that. You supply the logic; the platform executes it and lets you test it historically first.

If you have a risk profile, a sense of how much concentration and drawdown you can hold, and you want a documented method to translate that into a diversified portfolio and maintain it over time, Acubic is built for that. The method is published, the process is the same for every user, and the output is a specific, reviewable allocation before anything is executed.

Acubic does not promise or imply investment returns, and it does not claim the construction method will outperform any benchmark. It builds and maintains a risk-calibrated, diversified portfolio from your stated inputs. What the market does with that portfolio is outside the model.

If the job is building a risk-calibrated portfolio at your existing eToro or Trading 212 account, the Acubic portfolio builder runs the full construction process and shows the proposed allocation from your risk inputs before connecting anything.

Sources

1. Statistical Overfitting and Backtest Performance, David H. Bailey, Stephanie Ger, Marcos López de Prado, Alexander Sim and Kesheng Wu, 2014 (accessed 12 October 2026).
2. Optimal Versus Naive Diversification: How Inefficient Is the 1/N Portfolio Strategy?, Victor DeMiguel, Lorenzo Garlappi and Raman Uppal, Review of Financial Studies 22(5), 2009 (accessed 12 October 2026).