No-code Quant Trading

Inspiration

We are three data science students with strong coding and implementation skills, but limited background in quant trading. When we built our first model, we realized the hard part was not just writing code, but knowing which trading ideas were actually meaningful. That inspired us to build something that helps close the gap between strategy ideas and technical execution.

What it does

We built Simplified Quant Trading, a platform that makes quant trading more accessible and interactive. It combines a BTC prediction market trading strategy with a simple interface where users can adjust parameters and better understand how different choices affect the model’s behavior.

How we built it

We used the provided market data and built a strategy based on momentum, volatility-adjusted thresholds, order book imbalance, and arbitrage opportunities. We ran over 40 backtests and did hyperparameter tuning to improve performance and stability. We also built a website that lets users explore and customize the strategy more easily.

Challenges we ran into

Our biggest challenge was limited domain knowledge in trading. We could build and debug the system, but it was harder to know which signals were actually meaningful. We also struggled with organizing and comparing backtest results in a clean way. We also struggled with computing power when running a lot of backtesting. We end up testing in a smaller scale and also running more parallel

Accomplishments that we're proud of

We are proud that we built a working quant trading system despite starting with very little background in the field. We are also proud that we turned it into a more accessible product instead of just stopping at a technical model. Our project became a bridge between strategy ideas and implementation.

What we learned

We learned that quant trading is not just about coding, but also about market behavior, testing assumptions, and managing risk. We also learned that a simpler, more explainable system can sometimes be more valuable than a more complex one.

What's next for us

Next, we want to improve the model with stronger evaluation metrics and more cross-market signals. We also want to expand the website into a fuller no-code strategy builder so users can test and understand strategies more easily. If you want, I can make it even more Devpost-polished and punchier, like slightly more impressive without sounding fake.