QM028 Hands-on Lab: Adaptive Risk Control
Objective
Use a small synthetic monthly return sequence to compare a Core portfolio with a bounded adaptive risk-budget overlay. The exercise is intentionally simple: it demonstrates portfolio mapping and reward sensitivity, not a production RL result.
Run
python labs/python/qm028_rl_risk_control.pyWhat to inspect
- Which months are mapped to the defensive state?
- How does a 50% risk budget change return and drawdown relative to Core?
- What happens when the turnover penalty is increased?
- Why is a better outcome on this synthetic path not evidence that the state rule will generalize?
Interpretation boundary
The script uses a predetermined synthetic state sequence. It does not train or validate a deployable policy and makes no empirical market-performance claim.