Why We Publish Our Losing Markets

Almost every prediction service describes its accuracy. Very few publish it in a form anyone could check, and none that we are aware of publish the markets where their model is losing money. We do both, and the reasoning is straightforward.

A record that hides losses cannot be verified, and an unverifiable record is a marketing claim. Publishing every settled pick with the final score attached is the only version that lets a reader draw their own conclusion rather than accept ours. That includes the uncomfortable rows.

Our worst-performing market at the time of writing is running well below break-even, and it sits in the same public table as the best one. Anyone can download the whole thing as CSV under a CC BY 4.0 licence and recompute every number we quote. That is what a record means.

Key points

An unverifiable record is a marketing claim, not a record
Every settled pick is published with the final score
Losing markets appear in the same table as winning ones
Full dataset downloadable as CSV under CC BY 4.0

Frequently asked

Why publish losing markets?

Because a record that excludes them cannot be checked, and a record that cannot be checked is worthless as evidence.

Can I verify the results myself?

Yes. Every row includes the final score, so any pick can be checked against the actual match result.

Where is the data?

At our open data page, downloadable as CSV and JSON under a CC BY 4.0 licence.

Put it to use

Why We Publish Our Losing Markets

Key points

Frequently asked

Put it to use

Read next