A program that will take market data and then decide when to trade.
I created the program in 2020 during my gap year.
Here are the most important files in the program.
- The historical data handler
- Backtester (the engine of the program)
There are several modules:
- Data Handler
- Strategy
- Portfolio
- Risk Handler
- Broker
- Backtest
There is a detailed progression to each module.
The backtester runs on heartbeats. Each heartbeat, new data is made available.
The DataHandler is responsible for retrieving and making the new data available. First it instructs the Portfolio module to update the holdings with the new data. Then it calls Strategy module. The Strategy module is responsible for deciding whether to trade based on the newly available data.
If Strategy decides it wants to trade, then it calls the RiskHandler module. RiskHandler decides how much to trade (position sizing), and then calls the Broker module. The Broker module is responsible for making the trade online and updating the holdings in the Portfolio module.
Currently there are two types of data that the DataHandler can handle, SENTIMENT and BAR data. BAR data is just asset prices. SENTIMENT isn't completely implemented, but essentially it just scrapes twitter too see how people are feeling about an asset.