As mentioned in my last article talking about the Steam Hardware Survey, Valve sometimes adjusts the numbers a few days later. They have again and this time they adjusted the Linux market share on Steam down, not up.
Originally, the December market share for Linux on Steam was shown as 0.43%, but Valve have revised this down to 0.26%. So that's a decrease of 0.01 percentage points from November to December.
I'll be honest, I was rather surprised to see that the Linux market share on Steam had suddenly shot back up a bit, considering the Windows-only game, PLAYERUNKNOWN'S BATTLEGROUNDS is still pulling in a ton of people from the Asian market and is still breaking player records. In hindsight, it didn't really make sense for the Linux share to have risen like that. I still maintain my position that it most definitely will go back up, but that will be once the hype around PUBG fades.
Still, it's strange, as this is the biggest amount Valve has ever revised the survey numbers since I've been covering them. Well, as big as the percentages we're talking about are anyway…
So basically, nothing really changed between November and December. Here's to a strong and stable 2018.
Thanks for the news tip facePlanted.
Article taken from GamingOnLinux.com.
I'd just be curious how many other people read the blog despite not actually gaming on Linux. (In my case out of a mixture of interest in the technical aspects of Linux support, and in how the Linux-requirement seems to emphasize interesting/well-made indie games.)
Though I usually read the blog on iOS/Android.
Quoting: rat2000Isn't the methodology the survey that we participate at random times? I think it just checks then the OS Steam is running on and that's it.According to the last survey timestamps I've had a survey a few times, but I don't even know how it looks like. Although I vaguely remember something during the winex/cedega period.
https://www.reddit.com/r/linux_gaming/comments/7ail87/howto_get_access_to_the_steam_hardware_survey/
My surveydate on my steam machine was 2017-07-05, which of course is a lie...
Last edited by Ardje on 8 January 2018 at 9:20 am UTC
Quoting: BasianiAll WINE users, dual booters and VM users are counted as Windows users. So whats matter?Yeah, absolutely. I took part in the survey at least 5 times in the last 2-3 years. Each time, I just happened to be on Windows (once via Wine)...
lucifertdark Jan 8, 2018
Quoting: TheSHEEEPYeah, absolutely. I took part in the survey at least 5 times in the last 2-3 years. Each time, I just happened to be on Windows (once via Wine)...Then stop doing it, only do the survey when it pops up on Linux.
[email protected] Jan 8, 2018
BlackBloodRum Jan 8, 2018
Firstly, it can only count those who have steam accounts. Not all Linux users have steam.
Secondly its numbers are way out of sync with everywhere else... 0.2%? Almost all other places rate us at 1-2% range.
In addition the data is incomplete, 0.2% okay, but 0.2% of how many survayed?
You could come to the conclusion that perhaps many Linux users are not using steam for verious reasons, No games of interest? Don't like DRM? Simply don't play games?
Lets not also forget there are a huge chunk of Linux users who'd rather game on a console, and work on their tux. Thereby again, not using steam.
Basically all I'm saying is you cannot draw a conclusion on Linuxes marketshare with Steam data alone.
Last edited by BlackBloodRum on 8 January 2018 at 10:42 am UTC
Quoting: BlackBloodRumIn my opinion the steam survey is a highly unreliable way to count Linux marketshare (as a whole).
It doesn't count overall Linux market share, nor is it intended to do so. It counts STEAM USERS ON LINUX and nothing else. Important distinction which a lot of people seem to miss.
Instead, they need (and should be able to see) stats which shows user percentage on a set of games that actually have useable Linux ports. That would show the real numbers a Linux port can expect. Remember articles where game developers reported Linux user percentage like 3-5% or even more.
Quoting: GuestThe survey is pointless anyway since Valve does not mention the sample size.
Actually... If you have multiple incomplete studies, you can correlate them against each other and infer the estimated number of unknown cases. Something like this https://granta.com/violence-in-blue/
I think it should be possible to at least find the order of the sample size of the steam survey via other statistics. But it's definitely something for a statistician better than me.
Patreon
PayPal
See more from me