Trade Like It’s 2007

Inspiration

Crypto Markets are considered to be volatile. This opens opportunities for trading and profits. We aspire to create a profitable trading bot.

What it does

We have created a Trading Bot that creates predictions for the market and when they are considered to be favorable for a buy order, crypto buy order is places for the crypto exchange

How we built it

Laravel framework (PHP) was used as a base for the project. First we gathered historical data from external services - binance, glassnode and gathered data from there. Then we used gathered data to calculate indexes - combined RSI and combined TVP. Prediction is made based on the mentioned calculations. If prediction is favorable, a buy order is created in our database. The Application also gathers live data to continue with live predictions. To check profitability, a backtester was implemented that simulates order buy/sell in the past to calculate profitability.

Combined RSI is Relative Strength Index implementation with different importance placed on different time intervals. It sums Monthly RSI, Daily RSI and other time interval RSI where each RSI has different ratio in the decision.

TVP index is Total Transfer Volume in the past where the current Total Transfer Volume is lower than N number of previous Total Transfer Volumes in pecentage. It is used to indicate possible Volume reversals in the market. If the Transfer Volume is a lot lower than N previous volumes than next volumes might start to increase. Increase in Volume creates Volatiliy.

Challenges we ran into

We had to readjust our initial idea, taking into account the value we bring to the market, Correct use of a fitting algorithm and Reaching actual profitability (this is more of a core problem that the bot will have to deal with all the time).

Accomplishments that we're proud of

Building a working project which we can work with and improve upon in the future. We reached the goals we set for ourselves even though we had doubts as this was our first hackathon.

What we learned

How to use the available data to adjust the decisions the trading bot makes and that it can have many factors. We also explored Trading Bot building in general. The topic is much broader than we thought.

What's next for Trade Like It’s 2007

Further Ratio adjustment for Buy Signal algorithms is required as now the results are uncertan.

Currently our order places are done in our database system and are only considered suggestions. As we already have integration with Binance API, next step would be to add real order placing with Stop Limit.

As we already have implementation for Buy orders based on Low RSI, we could implement Sell Order in Margin markets based on High RSI.