Process Joey O'Brien's ORE sensitivity analysis post

Table of Contents

This page is a capture in the inbox bucket of the product backlog — a pre-sprint idea, not yet pulled into a sprint as a story.

1. What

Work through Joey O'Brien's blog post "Sensitivity Analysis in ORE" (https://obrienjoey.github.io/post/ore_sensitivity) and distil it into the knowledge base. The post walks the full sensitivity pipeline end to end on a single teaching trade (a 20-year EUR-ESTER OIS, 10M notional): the zero-domain config (simulation.xml 10-tenor grid, sensitivity.xml absolute 1bp shifts on DiscountCurve/EUR and IndexCurve/EUR-ESTER), reading sensitivity.csv zero deltas, adding a ParConversion block (OIS/IRS/DEP/FRA/TBS/XBS/FXF instruments, SingleCurve=true) that produces =parsensitivity.csv plus jacobi.csv=/=jacobi_inverse.csv, the Jacobian change of basis (\(\nabla_z V = J^T \nabla_c V\), inverted as \((J^{-1})^T\)), and verification: bumping the actual 15Y OIS market quote ±1bp and re-bootstrapping gives −7,195.31 by central difference vs −7,147.46 by the Jacobian par delta (0.67% gap, linearisation vs full non-linear recalibration). The takeaway: par conversion is accurate enough for hedge instructions, VaR and stress testing, and it is built only from ORE's own MarketRisk examples. Relevant to ORE Studio's sensitivity reporting support.

2. Why

ORE Studio does not yet expose sensitivity analysis; this post is a complete, verifiable worked example (config files, output tables, mathematics, independent verification) that would anchor a knowledge page and inform the product design. It is one of a series of ORE posts by the same author that are best processed together.

3. References

4. See also

Emacs 29.3 (Org mode 9.6.15)