The Bit Defender

We were engaged by a client based in London who was looking to create an automated system for trading crypto based derivates while limiting the potential downsides of single day loses and flash crashes by using automated hedging strategies. The client required both the ability to model different strategies in order to find a suitable strategy and continue to improve it, as well as bots to interact with trading platforms and carry it out instructions automatically.

We built out detailed models using Pandas and Jupyter notebooks and with using detailed ticker data provided were able to build models that would simulate the trading environment, including trading fees, slippage and the changing price of the underlining asset. We used this to run multiple models with multiple parameters and were able to find the most optimal strategy based on the criteria of maximizing RoMaD and the Sharpe Ratio.

Following this we built out the trading bots which were programmed with the most optimal models and configured with the outputted parameters from our findings. These bots would connect to the selected trading platform and continue to run these models in a completely hands off manner while offering the flexibility to change input parameters based on changing market conditions. Our final real life results were found to be within a 5% margin of correctness from our models.