-
General
- Document all the function input parameters
-
Dry testing
- In addition to running our strategies on past data, we want to be able to pull live data (syncing with Bitcoin trading APIs) and dry-run our algorithms, so users can see the performance of their own strategies live.
- Support for pulling data from multiple exchanges (but still running single-exchange strategies)
- Support for multi-exchange strategies
-
Better Strategies
- Auto-compare strategies by running them with the same seeding multiple times and comparing results
- Machine Learning, especially to predict prescient algorithms
- General framework for a ML training interface
- Clean up EMA code logic & optimize its parameters
- Think more about its logic -- maybe 'order_once' should take into consideration the trend_threshold as well; 'follow_trend': trend detection should be less sharp and more forgiving; maybe qty=-1 isn't ideal (maybe it should scale with the trend?)
- Higher-performing strategies
- General framework for risk management of strategies
-
Performance Visualization & Evaluation
- Ability to graph multiply strategies in the same graph (run multiple backtests for the same graph)
-
Live Testing
- Live connection to the exchanges, being able to run strategies real-time with real-money.