Currently focused on daily call contracts (the most liquid), I plan to build six models (calls/puts × daily/weekly/monthly expirations). The goal is to explore arbitrage strategies, mainly short straddles, by spotting overpricing in these derivatives.
Why Thorp-Kassouf? Practical, linear regression-based, and transferable to BTC options. Classical Black-Scholes assumptions may not hold for crypto, but this approach is simple yet powerful for prototyping.
Data: Historical BTC data from 2017-01-01, 5-min granularity, stored in MongoDB; processed with Python libraries (NumPy, Pandas, SciPy, Statsmodels).
Disclaimer: Not a Python expert; code evolved from a proof-of-concept. Can be optimized/refactored, focus is on modeling and insights for short-straddle strategies.
Next steps: Explore ways to “learn” when a contract fits the model.
GitHub: [https://github.com/dradicchi/kassouf-btc-options](https://github.com/dradicchi/kassouf-btc-options)
Kassouf paper: [An Econometric Model for Option Price with Implications for Investors' Expectations and Audacity](https://www.jstor.org/stable/1910443)
Feedback on modeling, reproducibility, or ideas for backtesting would be appreciated.