We built Logicull to find and exploit mispriced binary contracts on the world's most regulated prediction exchange. Along the way, we learned that the hardest problems aren't the ones you expect; and the solutions are rarely obvious from the outside.
This blog is where we share what we've learned.
What to Expect
We'll be writing about the intersection of finance, machine learning, and prediction markets. Topics will range from the deeply technical to the broadly strategic to even some advanced tips for prediction markets!:
-
Model Architecture — How we design, test, and iterate on our autonomous trading agents. From Nuthatch to Razorbill, every generation was a complete rethink.
-
Market Mechanics — The structure of binary contracts, how pricing inefficiencies emerge, and why regulated exchanges behave differently from crypto markets.
-
Engineering Decisions — The infrastructure choices that keep our systems running at scale — async execution, real-time data pipelines, and risk management.
-
Research Notes — Explorations into signal processing, regime detection, and the statistical methods that power our edge.
Why Now
Razorbill represents our most advanced model to date. With a five-tier classification system, stop-loss and recovery, and multi-market execution, we've reached a level of sophistication that's worth documenting.
The best time to start writing was when we launched Nuthatch. The second best time is now.
That quote sounds familiar right? We believe transparency builds trust. As we continue to push the boundaries of what autonomous agents can do in prediction markets, we want to bring our community along for the ride.
Stay Connected
New posts will appear here regularly. If you want to discuss anything we publish, reach out through our contact page.