A couple days ago our team developed a revolutionary bitcoin trading algorithm that significantly out-performed market returns - we were ecstatic.

We soon learned, however, that the market is very unfriendly to innovators, lacking an accessible means to trade and backtest trading algorithms. This is where K2 steps in.

What it does

K2 is a comprehensive backtesting platform for currency data, specializing in cryptocurrency, offering users the opportunity to create trading algorithms and simulate them on real-time and historical market data. We seek to equalize the playing field by enabling the general public to develop and test trading strategies. We created our own simple, Python-based API for users to use to develop their algorithm. They can then specify a trading interval, time frame, and currency symbol, and our backend visualizes the cumulative returns and generates financial metrics, quantifying the algorithm's performance against the market.

How I built it

K2 was built in JavaScript, jQuery, HTML5, CSS3, Python, AJAX, Node.js, Pandas, Numpy, Empyrical.

Challenges I ran into

Integrating the frontend with the backend proved quite difficult for our team. Our backtesting API is Python-based, so we had to execute Python code from the client-side and seamlessly incorporate the response into our frontend.

Accomplishments that I'm proud of

Successfully creating a financially sound backtesting platform complete with a slippage model and transaction fees that can simulate trading strategies on real-time market data.

What I learned

Methods to integrate client and server-side systems.

What's next for K2

A great idea is just that- an idea. We hope to publish, grow, and promote K2 through b2b sales, proprietary use, and customer outreach to truly bring wall-street to main-street.

Share this project: