Portfolio research guides

A curated set of practical guides for portfolio construction, risk profiling, and validation design. Built for real decision workflows, not classroom-only examples.

Each guide covers one stage of the same process, and they are meant to be read in order. Risk profiling comes first because every later constraint depends on it: how much drawdown a portfolio is expected to absorb, over what horizon, and which losses would force a change of plan. Those answers set the boundaries the optimizer works inside.

Portfolio optimization comes next. The interesting decisions are rarely about picking an objective function. They are about the constraints around it: position and sector limits, turnover, minimum trade sizes, and what to do when the mathematically optimal answer is one you would not actually hold. A guide that ignores those produces allocations nobody can implement.

Backtesting methodology comes last, because a backtest is only meaningful once you know what you were trying to build. Walk-forward validation, out-of-sample splits, and honest cost assumptions exist to answer a narrow question: would this construction have held up on data it never saw? That is a far weaker claim than a performance figure, and a far more useful one.