Derivasys research bundle: DRV-RV-2026-07-08-01

Question
Can a rough-Bergomi-style model calibrated to the medium and long expiries of one frozen Derivasys SVI surface provide a defensible scenario backbone, and what additional treatment is required at short tenors?

Contents
- reproducibility.json: structured study configuration, source identity, hashes and limitations.
- results.csv: the reported candidate comparison used by the public research note.

Source artifact
The working research notebook was notebooks/derivasys_svi_rough_vol_snapshot.ipynb in the local PDM project. Its SHA-256 digest is recorded in reproducibility.json. The notebook worktree was not itself a published repository at study time, so this static bundle is the maintained public artifact; it does not claim that the private market snapshot can be redistributed.

Data availability
The input was one complete SVI parameter snapshot received from the production Derivasys WebSocket at the recorded UTC time. Raw exchange books and the original transient message are not redistributed. The published results and figures are derived outputs.

Production status
Research prototype only. The selected smooth-premium candidate is a scenario base, not production calibration infrastructure or a live product claim.