About Predictarium

Predictarium is built by quants who've spent their careers pricing things for a living. Our backgrounds are in London's high-frequency trading scene and advanced STEM degrees from Oxford, and we've spent years trading and studying prediction markets, so this is the course we'd have wanted when we started out.

Who we are

We came to prediction markets from trading, and mostly from the high-frequency end of it in London, where you learn quickly that an edge is usually a few points of probability measured honestly over thousands of trades. Before and alongside that we studied at Oxford, and we've done academic research in machine learning, deep neural networks and deep learning, and in quantum computing. A fair amount of that feeds straight into what we build here, from the model behind the esports odds to the forecaster we're training for the API (which we won't release until it beats a simple rating model on matches it hasn't seen).

Why we built it

Most of what you'll read about prediction markets is either hype or a warning that nobody can win, and neither matches what we've seen in years of trading them. People do win, and the ones who keep winning tend to be a little more accurate than the price on markets they understand, careful about what fees and spreads take off, and sensible about how much they put on each time, so that a bad week doesn't finish them before their edge pays. We wanted somewhere that teaches exactly that in plain English, with real numbers you can check for yourself, and that lets you practise before any money's involved.

How we work

We check everything before it goes up. Facts come from papers, exchange rulebooks and fee schedules, which we link so you can read them yourself, and every number on the site has been worked out in code before it's published. The esports odds come from real match data, and the practice markets are clearly marked as simulated. If you spot something that looks wrong, we'd really like to hear about it at hello@predictarium.win, because that's how the course gets better.

What you'll find here

There's a course that starts from reading a price as a chance and works up to sizing, arbitrage and making markets, with an Ask AI button on every section if something doesn't click. You can trade the practice markets without spending anything, follow live esports odds built from real data, read our essays on the questions we kept running into, and, if you'd rather have the maths inside your own tools, use our API.