Construct a custom portfolio from EQS's full ranked candidate universe across large-, mid-, and small-cap sleeves — tuned to how you weigh yield, return, and volatility. Same subscription as Sector Leaders.
Included with the same Sector Leaders subscription — build and rebuild custom portfolios from the full candidate universe as often as you like. Cancel anytime.
Plain English works best — for example: “A 20-stock income portfolio tilted toward health care and staples, nothing in energy, and lower volatility than the market.” EQS drafts the holdings and weights from the current candidate universe, the sliders below adjust to describe what it built, and every ticker is re-checked against the same rules the manual builder uses. Fine-tune from there.
Rules-based construction, not a mean-variance optimizer — positions are scored and sized from conviction, yield, return, and volatility inputs within your sector-weight cap, not solved for an optimal risk/return tradeoff. When you build from a description, that text and the current candidate data are sent to an AI model that drafts a weighted holding list; the same rules then validate every ticker against the universe, drop duplicates, enforce the holding-count and sector limits, and rescale weights to 100%. The AI-drafted step is not deterministic — the same description can produce a somewhat different portfolio on a later run — and the model's judgment is not investment advice. The estimated portfolio volatility is an estimate, not a true representation of portfolio volatility: it starts from the weighted average of individual holdings' volatility, then applies a diversification adjustment based on an assumed 0.30 average correlation between holdings and how concentrated the book is — not a real covariance matrix, which we don't have data for. Actual portfolio-level volatility could run higher or lower than this estimate depending on how these specific holdings really move together. For research and informational purposes only; not personalized investment advice. Past performance of underlying names is not indicative of future results.