Scaredy-cat bot (team-040)

Inspiration

As Mathematics students, Optiver's challenge was the perfect fit for us, as it allowed us to put both our mathematical and programming skills to the test!

What it does

Powered by Optiver's platform, our algorithm models market dynamics in a simulated environment. It fetches real-time data, strategically adjusting positions to capitalise on potential market inefficiencies. By examining volume asymmetries and other traders' positions, our algorithm anticipates and reacts to market movements, determining optimal price points. Our focus was to react to market news and withdraw from trading during volatile market conditions to reduce risk exposure.

How we built it

Using Python, our algorithm interacts with Optiver's platform through an integrated API.

Challenges we ran into

Given that our goal was to exit asap when things get too heated, we found it difficult to constantly provide updated instrument prices as well as react to market news at the same time without slowing our algorithm down too much. We also faced a challenge when training our algorithm to interpret news feeds.

Accomplishments that we're proud of

We are quite pleased with our performance. We successfully managed to implement an exit strategy which reacts quickly to market news. Additionally, using text classification models, we managed to train our algorithm to read news feeds and understand which company it concerns, if it's good, neutral or bad news and how impactful the information will be.

What we learned

As Mathematics students we greatly improved our knowledge of market making as well as programming skills. We succesfully implemented text classification models for the first time as well as understanding multi-threading in Python.