Custom trading strategy implementation and backtesting. Ashrith Bandla, Alex Beloiu, George Fane, Shashank Kalluri

  • 11,588 Raised
  • 1 Judges


  • No tag


  • ETHChi for Good
  • Partner - Tastycrypto x Amberdata






Technologies Used: Amberdata, OpenAI ChatGPT 4, React, Python Backtrader and Flask

Description: This project aims to make crypto investing more accessible to people who are new to DeFi, or to traders who want to easily explore strategies.

Our app allows users to describe in plain language a trading strategy for backtesting. They don't even need to code it, because we use OpenAI to implement the strategy in Python. We then use Amberdata to retrieve daily data for the past five years and backtest the strategy, showing trades and final portfolio value today if the strategy had been executed.

Immediate improvements are to let the user tailor the timespan and frequency of historical price data, which is currently set to daily OHLCV for the past five years. Just to make things simple, we only retrieved the Bitcoin price feed, Additionally, we will let the user directly modify the Python trading strategy generated by OpenAI if they see issues.

In terms of longer term goals, now that the user can see the backtested performance of their custom strategy, we will create the functionality to automatically trade on that strategy. Our current plan is to programmatically spin up bots that transact on Uniswap. Further down the line, we will incorporate SocialFi, allowing other users to join in on a strategy for a fee.

By making crypto investing much easier, we can introduce much more capital into the blockchain ecosystem from people who are wary of the environment due to uncertainty. No matter where people are in their financial journey, they can improve their expertise and take control of their wealth with Reverberate.