Skip to content

Portfolio construction

Build a portfolio, not a pile of strategies.

The AI portfolio agent reviews candidates in the context of the existing book. Portfolio analytics compare return relationships, allocation choices and risk contributions using the available overlapping history.

Drawdown & exposure review

Inside portfolio construction.

Explore how this stage contributes to the AI trading strategy factory.

Check shared exposure

A correlation matrix reveals strategies that move together. Effective bets help distinguish the number of strategies from the amount of independent exposure.

Compare allocation schemes

Explore equal weighting, inverse-volatility weighting and risk parity. Each changes how the historical portfolio distributes capital and risk.

Review the whole book

Read portfolio Sharpe, drawdown, diversification and risk contributions together. Promotion also considers existing strategies and available paper capacity.

Consider the candidate alongside the book.

Portfolio analysis relies on enough overlapping returns. Missing evidence should stay visible instead of turning into a reassuring diversification score.

  1. Align histories

    Compare strategies over a shared observation window.

  2. Measure overlap

    Inspect correlations and effective independent bets.

  3. Compare weights

    Study allocation and risk contributions together.

  4. Review admission

    Combine the evidence with AI portfolio review and release checks.

More strategies can still mean the same bet.

Read the relationships before interpreting a combined return. Correlation and risk contribution describe different parts of the portfolio.

Compare independence with allocation

A and B in this illustration share much of their return behavior. Adding C changes the mix of exposures, but the weights and individual volatilities still determine how much risk each contributes.

Illustrative correlation matrix: A and B move closely together at 0.80. C has lower correlation with both at 0.10 and 0.20. These are synthetic values, not current portfolio results.

Different weights answer different questions.

The portfolio selector changes an analytical comparison. Choosing a scheme does not submit exchange orders or authorize live capital.

Scroll horizontally to see all columns.

Allocation schemes available in portfolio analysis
SchemeHow it allocatesInterpretation
Equal weightGives each strategy the same capital weight.Simple capital balance does not imply equal risk.
Inverse volatilityGives lower-volatility strategies more weight.Uses individual volatility; correlations still affect combined risk.
Risk paritySeeks to balance contributions to portfolio variance.Depends on estimated covariance and the available history.

Questions, answered.

More questions and answers
Where can I inspect my portfolio?

The Portfolio section in the signed-in workspace analyzes your graduated strategies. It needs sufficient overlapping return history to calculate meaningful portfolio statistics.

Does changing the allocation selector rebalance my account?

No. It recalculates the historical portfolio analysis. It does not change an exchange account, submit orders or authorize a live allocation.

What are effective bets?

Effective bets summarize how much independent variation is present in the correlation matrix. Several highly correlated strategies can behave much more like one bet than several independent bets.

Why can equal capital produce unequal risk?

Volatility and correlations determine how each strategy contributes to portfolio variance. A small capital allocation can still contribute a large share of risk.

Does portfolio review guarantee diversification?

No. It evaluates the available evidence. Correlations and losses can change in stressed conditions, so historical diversification is not a guarantee of future protection.

How does this relate to signal books?

Signal books combine forecasts within a strategy. Portfolio analysis combines the returns of multiple strategies. They address different levels of exposure and should be reviewed together.

Explore the connected workflow.

See how AI agents, testing engines and execution controls work together across the strategy factory.

Stratmill is an AI trading strategy platform, not financial advice or a broker. Backtest and paper results are hypothetical. Trading involves risk of loss.