What exactly is being priced?
A market starts with a precise contract. We examine event definitions, observation windows, settlement rules and differences between venues.
What would change your mind?
The world does not owe a model consistency. We study how information becomes a price, how a price becomes a position, and where that reasoning might break.
Evidence before conviction.
We investigate signals, pricing relationships and market behaviour. A useful result must survive scrutiny of its assumptions, data and execution costs.
From observation to execution.
Reliable data pipelines and observable systems support the research process. Automation creates more room for judgment, investigation and careful review.
Read the clock.
Information has a chronology. We care about when evidence became available, which decision it could have informed and how much uncertainty remained at that moment. Research should be examined with the information that could actually have been used.
Challenge the comparison.
Related contracts are useful only when their differences are understood. A shared headline can conceal different thresholds, observation windows or evidence sources. We work through those differences before reading a relationship into their prices.
Make tools accountable.
Code and AI can accelerate investigation, but speed only helps when results remain traceable. We want to understand the source, transformation and assumptions behind an output well enough to question it, reproduce it and decide whether to use it.
Research, in the open.
Explore our four organized and cleaned prediction-market datasets, publicly released on Hugging Face.