Quant trading summary
25 Oct 2024
https://blog.headlandstech.com/2017/08/
Companies are usually split into
- Strategy research: programming, stats, trading intuition
- Development: low level, networking, system architecture
- Ops
What kind of programs exist?
- Programs to parse exchange data, make trades, and emit orders
- Programs that support live trading
- Research stuff
What kind of strategies exist?
- Arbitrage:
- To be competitive here, you need scale in compute
- Faster telco speed, microwave towers!
- Queue position
- Market taking:
- Just help things be priced correctly, an example being recorrecting the market after a large buy order.
- Market making
- Profit from bid ask spread, help connect buyers and sellers.
- Model based:
- Volatility models
- Triangular arbitrage
- Implied chains
- Weighted constituent price calculations for ETFs
- Rule based
Market microstructure signals: are just signals produced from data feeds.
FPGA land
- Pretty much many places just have FPGAs to do the fast execution, and have a slower controller to provide the FPGA with conditions to execute on.
- This pushes execution speed down to less than a microsecond.