The Steam Labs features Valve tests out for its PC games store received Play Next as an experiment earlier this year. The feature was touted as a way to let players find a new game to play from their own library that they haven't yet given a chance to, all using machine learning. Now, Play Next has graduated from the experimental stage to being part of the Steam client thanks to positive feedback from users.
In the original announcement blog post, Valve said this solution will give out personalized suggestions as the algorithm tries to find titles a user will enjoy the most. The system is similar to another Steam Labs experiment, the Interactive Recommender, but instead of new games to purchase, this focuses on the already owned library of games.
Unlike the Steam Labs implementation, where Play Next shows three suggestions at a time with some additional details like trailers and tags, the Library version is displayed as a new shelf that can be moved around or removed altogether. Depending on the user's library of games that haven't been touched or have very little playtime, it can display quite a few games using the algorithm that tries to find an experience the user might like.
Including today's new addition, there are currently nine active experiments available for Steam users to check out in the Steam Labs section. Some of the most recent arrivals include the Steam News Hub and the updates to the store's search functions that allow pricing and tag-based filters.