QM008 Hands-on Lab

Recompute drawdown depth, underwater age, recovery, and censoring

QM008 · Hands-on Lab

Goal

Use a small wealth path to reproduce the state variables in the article:

running high-water mark;
peak-normalized drawdown;
signed underwater value;
current underwater age;
completed peak-to-recovery duration; and
a right-censored unrecovered episode.

Data

The input is data/qm008_wealth_path.csv. It contains 11 regular-grid observations and is deliberately designed to include exact equality recoveries and one incomplete final episode.

Python

python python/qm008_hands_on.py

The script uses only the Python standard library.

R

With base R available:

Rscript r/qm008_hands_on.R

Both implementations use the same QM008 convention: positive peak-normalized drawdown, equality recovery, elapsed peak-to-recovery intervals, and right-censoring when recovery has not occurred by sample end.

Try one change

Change the last wealth observation from 100 to 122 and rerun the code. The final episode should switch from censored to completed because re-touching the prior peak counts as recovery under the article’s convention.

Do not change the recovery condition from \(W_q \ge W_p\) to \(W_q > W_p\) in this exercise. That would be a different convention rather than a different observed path.