Automated backtesting platform for crypto trading – Case Study

Kukel Attila

2024-08-05

The crypto currency market is constantly changing and you need the right tools to trade effectively. Our automated backtest solution developed for our client helps you to do just that. The system we have developed allows users to collect and analyse real-time historical stock market data and optimise their trading strategies.

Historical data collection and storage

The script saves the historical data and the latest data after the candle closes by iterating over predefined crypto token pairs (e.g. BTCUSDT, ETHSOL) and time frames (e.g. 4h, 15m, 1m, 1s). The script retrieves and saves data by running in parallel on multiple threads, thus meeting high intensity write requests. A high-performance database system was used instead of traditional file-based or relational databases.

Trading strategy analysis on historical data

Data analysis was the biggest challenge of the project and the heart of the ecosystem. The analysis script is written in Python and the system is able to run the analysis on both CPU and GPU. The analyses look at various technical indicators and specific parameters to determine the best profit-generating values. We assisted in porting trading strategies provided by the client and provided training on the appropriate methods.

Generate trade signals

The third module of the ecosystem is the provision of real trading signals. The data is continuously saved and analysed, and the trading signals parameters are automatically refined based on the results of the analysis. This allows users to open or close positions in the crypto market in a timely manner.

Summary

Our automated backtesting platform takes crypto trading to a new level. The collection and analysis of real-time historical data allows users to optimise their trading strategies, increasing efficiency and profitability. Innovative database management solutions ensure fast and reliable operation, even with large amounts of data.

About the project

Throughout the project, we have successfully solved several challenges, including real-time processing of large amounts of data and GPU/CPU parallel execution capabilities. Based on customer satisfaction and successful improvements, we are proud of this innovative trading solution.

If you also have a software challenge for us and are inspired by this case study, feel free to contact us!

Join our newsletter!

If you liked our article, subscribe to our newsletter!