@alessandro (Here is a copy of my reply in another instance.)
Looks like the growth comes mostly from Windows (-0.56%) users switching to Linux (+0.52%). MacOS (+0.05%) users mostly seem just to upgrade MacOS and are mostly unaffected by the overall numbers. Inside of the Windows numbers, Windows 10 (-1.56%) users switching to either Windows 11 (+0.92%) or choosing an alternative platform (-0.56%). Numbers do not add up perfectly, because these statistics are estimation based on asking randomly a fraction of the user base.
There was some earlier article that a large portion of Linux users are using the Steam Deck, which is Linux-based.
googles
Not the article I was thinking of, but makes the point:
https://www.gamingonlinux.com/2023/08/linux-overtakes-macos-users-on-steam-thanks-to-steam-deck/
42% of Linux Steam users are Steam Deck users.
It’s very good news for Valve, which is seeing their platform do well. And it’s good news for users of other Linux distros in the sense that compatibility with Valve’s Linux probably helps ensure support for them.
But it doesn’t mean, I think, so much that lots of Windows users are moving to Linux unless they are abandoning their desktop for a Steam Deck. More that there are likely people using both a Linux-based device – Steam Deck – and a device running another OS.
To be fair, it’s still good for Linux because in this case, software distributed to Steamdeck through Flatpak or Steam can largely be used by other distros as well, which can make more conventional desktops more enticing in the near future.
Hey! I am one of the 2%. I left Windows around 3 years ago and love gaming on my Linux PC.
@alessandro But the Steam Deck does not get hardware survey in Gaming Mode. It only has a chance to get a survey, if the user logs into Desktop mode and then have to stay there for long time in order to get a chance. So I don’t know how the statistics are calculated here.
Same perks and disadvantages apply on all platforms Steam collect data from. My guess Valve do apply some weird AI algorithm to determine missing data. IE (just speculation) if there are Steam Decks that never go in Desktop Mode the algorithm pushes itself more widely in the array of the SDs that do.
One thing worth to remember, Valve always knows which binary executable connect to their server (and today very few use Steam’s windows .exe through WINE), the survey is just them asking permission to eventually use your data: is not a commitment, they can fix/screw the statistical data anyway they see fit… I think mostly to fight those who cheat with the stats… Chinese language jumped few time over English (making English a secondary language on Steam) and Valve “fixed” those published data too!
@alessandro Valve knows what binary it is, but that does not mean Valve would know what distribution it is. There are no hidden data sent to Valve. On the other hand, Valve knows exactly how many Steam Decks are sold. This fact can be used in the statistics. The question is, if these statistics are ONLY about the survey results or if such information is included as well?
deleted by creator