Prediction market — order-matching engine.

Solution
Scale
Effort
3 FTEon average
Together for
3 yearssince 2023
Status
Active

01 · The client

A company that built a market where people take positions on future events, aimed at a professional audience. A new product, in a category that's only existed for a few years.

Type
Fintech startup
Region
Europe
People
Under 20
Stage
New product

02 · The trigger

They had the business model and no one who knew how to build the machine that runs it.

There are no specialized vendors to turn to in this category: the few who've built one built it for themselves.

03 · The technical problem

Orders that have to be processed in the exact sequence they arrive, in milliseconds, while each one's exposure calculation changes at the very moment it's being computed. All of it with a load that sits nearly idle for days and then all arrives in the final minutes before an event closes.

04 · What we did

We built the order-matching engine and the exposure and fee logic, with full tracking of every operation: months later, it must be possible to reconstruct why something happened the way it did. And we changed how we worked together along the way. The initial scope had been defined as if anyone knew what to build, but in a category with no precedent that precision is a fiction: decisions get made as the problem is discovered. We moved to a model where you agree on the commitment, not the list, and it was better for both sides.

Orders−7 dayscloseLast 30 minutes−30 min−10 minclose
Over seven days the peak is a vertical line. Zoomed in, it is the minute in which the engine has to keep the exact order of every trade.

05 · Where things stand now

Project ongoing.

Lane: Scale ·  Vertical: Prediction Markets

Write to us