Model M is a cross-sectional equity factor model that outputs target weights for a long-short book. A researcher has already assembled a Quant Research Dossier recommending a bounded paper-trading trial. This topic is what happens before independent validation can honestly begin: can a second person rerun the dossier, can a reviewer block silent lookahead, do numerical identities hold under stated tolerances, and is the eleven-item documentation package complete enough that misuse is less likely than silence?
By the end of this note you should be able to:
- judge whether a research claim about M is replayable from a run record alone;
- write review comments that stop validation-blocking defects before they reach a formal review;
- choose which identities deserve tests and which tolerances are economic rather than accidental;
- map M onto the documentation headings that model risk management already requires, especially limitations and restrictions of use.
In this lesson
- Contents
- The big picture
The full lesson is for members.
Read the full lessonNew to SelfMadeQuant.com? See membership options or browse the Professional Quant Practice syllabus.