Optiver Trading Challenge

Final Run

We, represented as "Fill Me Maybe", built a sophisticated algorithmic trading system that orchestrates three competing strategies—Market Making, Statistical Arbitrage, and Sentiment Analysis—across dual-listed stocks with intelligent priority management and thread-safe coordination. The core challenge was managing multiple strategies on the same instruments while respecting a global rate limit of 25 requests/second and ensuring high-priority sentiment signals could preempt lower-priority strategies without conflicts.

We learned that concurrency primitives (locks, events, queues) are essential for thread-safe trading systems, financial math details matter (sample vs population std, microprice vs midprice), and modern LLMs can extract structured trading signals from unstructured text with proper prompt engineering. The system successfully runs 15 concurrent threads (3 strategies × 5 stocks) with ~20 req/s throughput, demonstrating production-grade coordination of competing strategies under real-world constraints.