Valve has introduced a new Steam Labs Experiment, which is now live on both the web site and in the Windows tray accessory app. Steam Labs Experiment 008: Play Next utilises Valve's machine learning algorithms to prompt you with suggestions about what unplayed games from your library to dive into.
I'm sure many HEXUS regulars will have quite extensive Steam games libraries that get added to more often than played through. That is certainly the case in Casa Tyson where bargains and bundles get added to the Steam library at a much faster rate than free time becomes available (also I am reluctant to stray from trusty classics which I know are both highly absorbing and guaranteed fun).
When you are looking to buy new games, or simply browse the Steam Store, the site/app uses Valve's Interactive Recommender tech to customise what you see. It utilises machine learning to look at games you seem to enjoy playing and makes personalised recommendations from the whole store. You can head to that link and customise how Interactive Recommender works with sliders, tags, and other settings.
Now Valve has taken this tech and applied it to the unplayed games in your personal library. You will note that the recommendations might include remasters of games you have played thoroughly but it's easy to move through the top picks or refresh the list to ponder more carefully on your next time-sapping computer game favourite. To make the decision process even easier the Play Next page includes various tags appropriate to the unplayed game and you can simply mouse-over the large thumbnail image to check out a 'microtrailer' for the action found within.
As might be obvious from the fact that this is an 'Experiment', the system is a work in progress and may sometimes come up with unexpected results.