Mission
Can you predict the unpredictable?
The idea behind Vatex, and the bet we’re making.
The method
From raw data to a forecast.
Compress
We gather broad real-world data about an event and squeeze it down to the part that actually tells you something.
Model
We build a statistical picture of how events like this tend to play out — what usually happens, and when.
Distill
We turn that picture into storylines — the handful of ways things could plausibly unfold.
Simulate
An AI plays those storylines forward, many times over. The result isn’t one guess but a spread of outcomes, each with odds.
The hypothesis
The bet we’re making.
Intuition says random events can’t be predicted — that’s what random means.
For years, computers were hopeless at predicting what a person would say next — language seemed too messy, too human. Then, once the models got big enough, that changed: they became the assistants everyone uses today.
Much of what looks random may just be a pattern we haven’t found yet. That’s our bet. Feed a model the right data — enough to cover most of the ways a situation can unfold — and it may find an edge where intuition says none exists.
This is a hypothesis, not a result. This site exists to test it in public: every forecast sealed before the event, scored after, misses kept next to hits. If we’re wrong, the record will show it.
What it could be good for
If the bet pays off.
Guess first, verify second. Many hard problems are still solved by trial and error. A model that offers a likely answer up front flips that: check that answer first, search only if it’s wrong. Wherever checking is cheaper than searching, even occasional hits save enormous work.
Predict forwards — and reason backwards. A model that captures the current state of something — a match, a market, a company — can run it forward to say what’s likely next. It could also run backwards, filling in earlier events it was never shown. Predicting something and understanding it become the same skill.
One engine, many industries. We’re building for prediction in general, not one specialty. Fields that run on uncertainty — finance, healthcare, insurance — could all draw on the same engine.