The Royal Trader

• 

  Optibook: A Stimulated Stock Market

Team: 057 Filename: ~/environment/sentiment_analysis/sol_notebook_final.ipynb

Inspiration

Optiver's challenge was a great match for us, as we were looking for a challenge to gain a deeper insight into algorithmic trading and test both our quantitative and programming skills.

What it does

We built a strategy to integrate sentiment analysis into an existing market-making strategy.

How we built it

There were 2 main layers to our strategy:

• Market making: This was already provided to us. However, we decided to constrain it to only the instruments that did not show positive or negative sentiments. If that was not the case, we decided to execute the sentiment analysis strategy
• Sentiment analysis: An all-in approach was used here. If the sentiment, for instance, was classified as positive, we decided to first delete all the existing asks, and buy the max limit of stocks of the respective instrument. We would then wait 25s, and sell all of those stocks. Once the order information is obtained in the main loop every 2s, we decided to create a separate thread that will execute the bid and ask for orders independently.

Challenges we ran into

Understanding the market-making strategy and definitely debugging the algorithm

Accomplishments that we're proud of

• Coming up with a unique and profitable strategy in just 2 days
• Being part of hackaTUM2023

What we learned

This was a great opportunity for us to learn about market-making, as we were quite new to this domain. We were successfully able to build a strategy for sentiment analysis and incorporate it to improve the existing trading bot.

What's next for The Royal Trader

We would like to further optimize our strategy. We would like to further experiment with the classifiers for sentiment analysis, potentially train it to social_feeds data set, as well as improve its classification time.

Built With

• amazon-web-services
• python