A new experimental feature has been revealed for Steam. This is designed to help you find suitable games faster.
What it’s about: For its distribution platform Steam, the company Valve has provided a new technique. This aims to make game recommendations more accurately tailored to the individual user. It is referred to as the “Interactive Recommender” and relies on a “neural network model”.
What’s new about the system: Until now, game recommendations were mainly based on user reviews and tags. The new model generates suggestions in real time based on striking data, such as playtime history, and has been trained on data from millions of Steam users and gaming sessions.
Game information, such as genre or price, is no longer considered.
What do I, as a player, gain from the innovation? The system will be highly customizable. You as a user will have the option to adjust the new recommender through various parameters. For example, only newer games can be suggested on request, or niche titles from the vast Steam catalog can be explicitly searched.
If the system is successful, it can save a significant amount of time in searching for titles and expand your library with unexpected games.
What does Steam gain from the new feature? On one hand, the new system promises an improved user experience in the form of more precision in searching for new titles. This could be a long-term reason for players to use the store.
On the other hand, the new function also brings benefits for developers.
They no longer have to optimize their games according to the algorithm used by conventional search models.
According to the visionaries at Steam, developers will be able to focus entirely on fun gameplay without worrying about which tags and metadata might influence the success of their product. Smaller titles from indie developers could particularly benefit from the new feature.
How can I try the new model? You can find and try the new recommender in the Steam Laboratories within the ongoing experiments.
The recommendation didn’t fit you? No worries, here you can find a guide on how to return the mispurchase:
What do you think of the new feature and have you tested it yourself? Let us know in the comments.
