Tutorials
End-to-end walkthroughs — Heston calibration, fBm Hurst estimation, vol-surface from quotes, AI surrogate training, pairs, risk, execution, interop.
Tutorials
Long-form, end-to-end walkthroughs. Each tutorial builds one concrete artefact (a calibrated model, a vol surface, a backtested signal) from scratch, with full code in both Rust and Python.
Planned tutorials
| Tutorial | What you build |
|---|---|
| Heston calibration | Fit Heston to a market vol surface, generate Greeks |
| fBm Hurst estimation | Simulate fBM, recover H via Fukasawa |
| Vol surface from quotes | OTM quotes → IV grid → SVI/SSVI fit → arbitrage check |
| AI surrogate training | Train a Heston NN surrogate from scratch |
| Pairs trading | Cointegration test → hedge ratio → z-score signal |
| Risk pipeline | VaR / CVaR / drawdown over a portfolio |
| Microstructure execution | Almgren-Chriss optimal execution on a synthetic LOB |
| Python interop | numpy → stochastic_rs → numpy round-trip in a notebook |
The tutorials are filled in incrementally. While they land, use the Quickstart plus the per-section example blocks under Processes, Distributions, Stats, and Quant.
Format
Each tutorial follows a consistent structure:
- What you'll build — a screenshot or a representative numerical output up front, so you know if the tutorial is for you.
- Prerequisites — sub-crates, Cargo features, Python packages.
- Setup — the boilerplate.
- Steps 1 … N — each ≤ 30 lines of code, with prose explaining the why.
- Result — numerical output, plot, or both.
- Where to go next — three cross-links into the catalog.
Python bindings
stochastic-rs-py — Python coverage for distributions, processes, pricers, calibrators, copulas, stats. NumPy in / out, 210 entries at v2.0.
Benchmarks
Criterion bench numbers — FGN CPU vs CUDA, distribution sampling speedups, and the all-backends matrix (CPU / cubecl / Metal / Accelerate).