Portfolio Construction

A single tradeable signal is only the starting point. The harder problem is assembling many of them — most individually weak, often correlated — into a portfolio whose risk-adjusted return survives costs and live trading. This topic covers that assembly problem: how signals combine, how much capital each should carry, and how to keep the whole book from concentrating risk you didn't intend.

The articles here move from diagnosis to construction to control. Attribution and correlation analysis tell you what you actually own — which signals are driving P&L and where their bets overlap. Position sizing and multi-signal optimization decide how to allocate across them, weighing each signal's standalone edge against its marginal contribution to the portfolio. Risk management closes the loop, defining the constraints and drawdown controls that the allocation has to respect.

Read together, they answer one question: given a set of imperfect signals, how do you weight and constrain them so the portfolio behaves better than any single component? Start with attribution to understand your exposures, then use correlation and optimization to allocate, and let the risk framework bound the result.