Financial prediction markets.

Prediction markets grew up on questions that end in yes or no. Money questions end in a number, and that one difference decides most of how the market has to be built.

What is a financial prediction market?

A market where the thing being traded is a published financial number and the payout depends on where that number finishes. Inflation, a policy rate, credit stress, a benchmark return. The contract settles by reading the number, not by judging whether an event happened.

Why does finance need a different contract?

Because none of its questions have a yes or no answer.

Ask whether a team won and the world hands you a clean yes. Ask how much inflation ran and the world hands you 3.1 percent. To put that in a yes or no contract somebody has to invent a threshold, and from then on the threshold is doing the work instead of the forecast.

What breaks when you force a number into a yes or no?

Two things. Being nearly right stops paying, because a threshold has no middle, so a call that was correct about direction can settle at zero. And expressing a real view takes several contracts at once, each with its own order book, which thins out the depth behind every one of them.

Both costs are covered in scalar vs binary prediction markets.

What does a market on a number look like instead?

The payout slides with the finish. A share pays a fraction of a dollar equal to how far up the range the number landed, so a forecast that is close pays close to full and one that misses badly pays little. No threshold sits in the middle deciding who was right.

What has to be true for one to work?

The number has to be published by somebody other than the venue, on a schedule, under a stated methodology. That is what lets settlement be a rule instead of a judgement, and it is the difference between a market a desk can hold and one it cannot.

How that resolution works is in prediction market resolution, and a worked example follows one trade through to settlement.