Steam gets a new 'Deep Dive' discovery tool and Community Recommendations

(Image credit: Future)

More work is afoot in the Steam Labs, the not-so-secret facility where Valve tinkers on new Steam features that could benefit from hand-on use but aren't quite ready for prime time. Experiment number five is called Deep Dive, and is actually based on a project called Diving Bell released in July by indie developer and occasional PC Gamer contributor Lars Doucet.

Diving Bell was designed as a new discovery tool for Steam, similar to the Interactive Recommender but with an "entirely opposite" approach. Instead of coming up with recommendations based on your Steam history, it begins with a game you specify and then uses "dumb algorithms" to come up with a group of recommended games based on preset criteria.

"Recommender is like a sommelier that uses their expertise to suggest a wine that pairs well with the courses you've already chosen," Doucet wrote about the project, while Diving Bell is more like a "Wikipedia binge." 

"You start with some topic, then you click on links within that topic that seem interesting, and before you know it you find yourself following some totally weird but fascinating bunny trail you never expected you'd go down."

Valve worked with Doucet to bring Diving Bell to Steam as Deep Dive, which has a similar layout but "now directly leverages Steam APIs to serve up recommendations and related game information." For any selected game, Deep Dive will recommend "very similar," "somewhat similar," and "similar gems," which Valve described as "little-known but well-loved" games that match the recommendation criteria.

A second new experiment, Community Recommendations, is a new category of recommended games that highlights games that are being actively recommended by Steam users. It lists games based on daily recommendations sorted by their helpfulness to users, which can be extended out to weekly or monthly reviews. Minimum playtimes on user reviews can be set to eliminate the impact of ten-minute rage reviews, and filtering by tags will also be supported.

The Interactive Recommender experiment has also grown a little, and now has its own place of honor on the Steam Store home page. The recommender will serve up suggestions based on your recent gameplay, and also make it a little more convenient to blast into the Steam Labs page so you can tweak it up as you like.

As always, these are "experiments" and thus may not work exactly, or at all, as intended. If you have thoughts on any of them, or what you'd like to see Valve fiddle with next, there's a Steam Labs discussion group.

Andy Chalk

Andy has been gaming on PCs from the very beginning, starting as a youngster with text adventures and primitive action games on a cassette-based TRS80. From there he graduated to the glory days of Sierra Online adventures and Microprose sims, ran a local BBS, learned how to build PCs, and developed a longstanding love of RPGs, immersive sims, and shooters. He began writing videogame news in 2007 for The Escapist and somehow managed to avoid getting fired until 2014, when he joined the storied ranks of PC Gamer. He covers all aspects of the industry, from new game announcements and patch notes to legal disputes, Twitch beefs, esports, and Henry Cavill. Lots of Henry Cavill.