An engine is useful only when its numerical output can be reconciled with a defined economic model. This final topic turns the programme's identities into evidence: exact sign checks, zero-loss limits, benchmark prices, paired sensitivities and an explicit boundary on unsupported extensions.
By the end of this note you should be able to:
- distinguish an implementation defect from numerical uncertainty and model limitation;
- derive fixed-model recovery sensitivity;
- specify path-count, grid and conditional-valuation convergence checks;
- explain what additional resources a full xVA extension requires.
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 Counterparty Credit Risk and CVA syllabus.