Valve launches a new "Interactive Recommender" that generates personalized lists of Steam games that users might like based on what they have played before. It does not replace the platform's existing recommendation systems, but it should give users another way to navigate Steam & # 39; s huge backlog in games – especially smaller titles that are often lost in the mix.
The Interactive recommendation uses machine learning, which Valve notes & # 39; all cool kids doing & # 39 ;, for what sounds like a collaborative filter system. Instead of focusing on tags or metadata, "it looks at which games you play and which games other people play, and then makes informed suggestions based on the decisions of other people playing games on Steam. The idea is that as players with broadly similar playing habits also tending to play another game that you have not yet tried, then that game is probably a good recommendation for you. "
Once that is done, users can refine the recommendations with tags and sliders. So you can only search for categories such as & # 39; Massively Multiplayer & # 39; or & # 39; Indie & # 39 ;. You can only record newer releases (the slider moves from 10 years to six months) or you can weight the results from & # 39; popular & # 39; to & # 39; niche & # 39 ;. The latter function works in the same way as the Last.fm music recommendation system, which the Interactive Recommender generally resembles.
Valve describes this section as "a very effective way to find hidden gems." My niche results were not unbelievably obscure; Steam recommended Void Bastards and Layers of fear 2, for example, among many other titles that have received the attention of regular games. But at least most of the games I hadn't paid much attention to and probably didn't want to hunt. And it was actually things that seemed interesting to me – unlike most games in my normal Steam discovery queue.
The biggest disadvantage is that you cannot remove items from the recommendation list so far, even if you have already added them to your wish list or have chosen to & # 39; ignore them & # 39; on the store page. So unless you buy a lot of the games – and buy specifically on Steam, so they are in your library – your returns may fall over time. This limitation also makes the results for "Popular" much less useful because mine contained multiple games that I already own in other stores and platforms.
Valve notes that the Interactive Recommender also doesn't help you find new games because no one is playing them – although it supposedly picks up releases "with just a few days of data." That is part of the reason why Valve does not change the existing. recommendation options based on different models and data. You will still see the same new, popular and popular games on the main page and you can still browse recommended tags, plus all other composite selections and lists that Valve offers.
The recommender is part of a new initiative called Steam Labs, where Valve lets users view the experimental functions. (Do not confuse with The lab, a series of experiments with virtual reality made by Valve, or Aperture Hand Lab, a demo for the Valve Index VR headset.) Steam Labs now offers two other projects: Micro Trailers, a large folder with six-second game trailers, and The Automated Show, a half-hour video compilation of new Steam game trailers. Valve also refers to an intriguing project with the code name "Organize your Steam Library with Morse Code", although unfortunately it does not appear.
While Valve links these functions to its users, the Interactive Recommender seems to be a nod to smaller developers who are frustrated by the storefront. The recent Steam summer sale from Valve was reportedly a decline for many indie makers – partly because of some confusing competition rules that encourage users to remove games from their Steam wishlists, and partly because developers felt that Steam promoted large mainstream games at the expense of more niche titles. Valve also promises that the Recommender will not force developers to play the system with specific tags, prices or advertising styles, but without knowing more about how it works, it is difficult to exclude users from discovering peculiarities in the engine.
Valve tends to push mechanical, data-based solutions to Steam Store issues, sometimes in situations where it seems inappropriate. (See for example his review bombing charts.) But here is & # 39; more data & # 39; exactly what people need to go beyond Steam & # 39; s front page – and remember that there are still a ridiculous amount of cool stuff on the platform.