Editorial policy
Umpiry News publishes tournament previews and reviews built on Umpiry’s tennis data. This page explains how each article is made, who approves it and how it is corrected.
How an article is made
Every article starts from a fixed set of facts taken from Umpiry’s tennis database at a stated time: results, scores, draws, rankings and the outputs of Umpiry’s models. The “Data and sources” block at the end of each article names that time.
An AI language model drafts the text from those facts only. It is told not to add facts of its own and not to guess.
An automatic validator then checks the draft against the same database. Every number, player, score, round and date has to match a fact, and wording that claims more than the data shows is refused. A draft that fails is sent back for a new draft or is not published.
An Umpiry editor reads each article and decides whether it is published. Nothing is published automatically.
Models and probabilities
Title chances, win probabilities and projections in an article come from Umpiry’s statistical models, and the article says so. They describe likelihoods, not certainties. They are not betting tips or advice. How the models work is explained in the Methodology.
Sources and photos
Match data come from multiple sources and Umpiry’s own processing. Each photo carries its author and licence. An illustration is labelled as one, and an archive photo from another event is labelled “Archive”.
Corrections and updates
When an article contains an error, we correct it and add a dated note under “Corrections” at the end of the article. The updated date changes only with a change to the substance of the article.
To report an error, write to us through the Contact page.
Who we are
Articles are credited to “Umpiry Editorial”, the editorial desk of Umpiry. More about Umpiry is on the About page.