You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
34 lines
1.4 KiB
34 lines
1.4 KiB
# Prediction Market Arbitrage
|
|
|
|
Supports Kalshi prediction markets, Nadex binary options, and CME event contracts (through tradovate client).
|
|
Draws basis future pricing from yahoo finance.
|
|
|
|
Currently has code for nasdaq and s&p pairs, but can easily support any contract traded on these platforms.
|
|
|
|
## Setup
|
|
Provide auth data in `*.auth` files in json format.
|
|
Relies on python venv with `requirements.txt`.
|
|
If you're going to use nadex, install node and `package.json` for lightstreamer.
|
|
|
|
The pair you're considering and the minimum spread to execute a trade is controlled in the runner files
|
|
e.g. `nasdaqdaily.py` or `spdaily.py`.
|
|
|
|
## Running
|
|
Just run `python nasdaqdaily.py` or `python spdaily.py` etc. You can provide "demo" as a sys arg to run in the sandbox environment (for
|
|
the trading platforms that support this).
|
|
|
|
Log data, of good spreads found and trades executed, are outputted to stdout, so you should run this on
|
|
tee (into `systemd-cat` if you're on the hetzner server so we can export from journalctl).
|
|
|
|
Login sessions are outputted into `*.session` files, which can be cleaned up using `./clean`.
|
|
|
|
## Testing
|
|
|
|
Run `pytest -rA`.
|
|
|
|
## Todo
|
|
- Add fidelity, and equities vs futures (this requires non-piecewise math solveer)
|
|
- Reverse engineer and reimplement lightstreamer in python
|
|
- Rigorously model execution risk better than just a minimum spread (look into Almgren-Chriss mean-variance and LVaR models)
|
|
- Refactor source code layout
|