The latest Steam Hardware & Software Survey is out now for September 2022, with it showing that clearly the Steam Deck with SteamOS is the most popular way now to game on Linux.
With the latest numbers though, there is a dip in the overall Linux user share with it now at 1.23%. A drop overall of 0.04 percentage points which some may attribute to just noise but that's what the survey says. Historical data can be found on our Steam Tracker like the below graph snapshot:
As for the most popular Linux distributions, it's now pretty clear that SteamOS is on top:
- SteamOS Holo 64 bit 17.04% +3.35%
- Ubuntu 22.04.1 LTS 64 bit 12.48% +4.28%
- Arch Linux 64 bit 10.57% -0.32%
- Manjaro Linux 64 bit 7.80% -1.13%
- Flatpak runtime 64 bit (multi-distro) 4.81% +4.81%
- Pop!_OS 22.04 LTS 64 bit 4.31% -0.15%
- Other 42.99% +5.10%
I'm sure some will argue about combining distributions together, like Arch and Manjaro that would put them on top but really they're quite different. Just like how SteamOS has its own update schedule, so does Manjaro they're just based-on Arch. A bit different for Pop and Ubuntu, which share the same base directly.
As for devices, Steam Deck is clearly top for Linux with it's GPU being used by 17.07% putting it at the top for Linux players. That's an +3.35% increase compared with last month too, so the lead is only continuing to grow. Once again though, it's still not even showing up in the Windows stats meaning very few overall people are using Windows on the Steam Deck.
It's a bit like that "PC gaming is dead! Mobile is the future" nonsense that crops up every now and again - with PC gaming being in better shape than ever and with more games than ever.
Just because some share shrinks (or grows), doesn't mean there are actually less (or more) absolute users.
You only get this "very rough ballpark" number when Valve shares some actual user numbers once in a blue moon.
As far as I know, the Steam Deck and only the Steam Deck has got the "AMD Custom GPU 0405". This comes up in the general GPU list with 0.21% of all systems. Linux is installed on 1.23% of all systems, and when filtering by Linux, we can see that 17.07% of the Linux installations are using said GPU. And guess what 1.23% * 17.07% is? Right, 0.21%.
Last edited by Eike on 3 October 2022 at 2:26 pm UTC
QuoteThat's an 3.35% increase
Do you mean [percentage] points?
I've been looking at GoL's survey. OSS GPU drivers surpassed proprietary ones recently (AMD pretty close to overtaking nvidia). Interesting trends all around, included for Wayland adoption, which seems to be picking up (btw, there's no "window managers" on Wayland. Not yet.)
I'm wondering if you could do monthly "one-off" surveys in the reminder post. Maybe also accept anonymous input (counted separately). The topic could be quite random. "Will you get a steamdeck", "have you tried VR", "What language do you have Steam set to", etc.
Quoting: EikeNobody's installing Windows on the Steam Deck.
Speak for yourself.
Quoting: SikSlayerQuoting: EikeNobody's installing Windows on the Steam Deck.
Speak for yourself.
I showed the numbers. Got different ones?
(Obviously, there might be a handful of individuals that do install Windows on Steam Deck. These will not show up in my numbers as their count is insignificant. I expect everybody reading here to understand that without mentioning.)
Last edited by Eike on 3 October 2022 at 4:13 pm UTC
Here's an alternative way of looking at the numbers:
September 2019 - 0.83%
September 2020 - 0.94% (+0.11%)
September 2021 - 1.05% (+0.11%)
September 2022 - 1.23% (+0.18%)
Quoting: TheSHEEEPWith these, I always wish Valve would give out absolute numbers instead of percentages.
It's a bit like that "PC gaming is dead! Mobile is the future" nonsense that crops up every now and again - with PC gaming being in better shape than ever and with more games than ever.
Just because some share shrinks (or grows), doesn't mean there are actually less (or more) absolute users.
You only get this "very rough ballpark" number when Valve shares some actual user numbers once in a blue moon.
I don't think it is good idea. Valve dont have actual data about all, because every month only subset of their user is in survey. So % is better solution, because if they write absolute number, you expect real value.
Side note: -0.04% is so small number that can be be statistical error and reality can be positive change just because some bias in selection process or small sample size...
Quoting: ArtenValve dont have actual data about all, because every month only subset of their user is in survey.
As I have been complaining forever, Valve could easily take the data from live Steam client connections. They really have daily information of the base OS for every single user. It wouldn't have the hardware detail, but the OS statistics would be a 100% sample of logged-in users. Not sure why they don't.
Quoting: CyborgZetaInteresting that Ubuntu's back in 2nd place, and both Ubuntu 20.04 and Linux Mint have disappeared entirely.Ubuntu was in first place last month. SteamOS has now overtaken it by virtue of shipping a shedload of units. Ubuntu 20.04 has dropped as people have upgraded to 22.04, and now both that and Mint are low enough proportions that they're in that huge 43% Other.
Quoting: no_information_hereAs I have been complaining forever, Valve could easily take the data from live Steam client connections. They really have daily information of the base OS for every single user. It wouldn't have the hardware detail, but the OS statistics would be a 100% sample of logged-in users. Not sure why they don't.Valve know what OS each customer is using. Sharing that information with third parties without specific consent is a privacy violation.
Quoting: CyborgZetaInteresting that Ubuntu's back in 2nd place, and both Ubuntu 20.04 and Linux Mint have disappeared entirely.
Ubuntu 20.04 and Linux mint are no longer in the top 6 distros.
All distros outside of the top six and lumped in with "Other 42.99%" And that group grew by +5.10% in the last month.
Market share increases: "There are dozens of us! DOOZEENS!! Year of the Linux Desktop TM is coming" (been there myself, done that)
Market share decreases: "Minor decrease, statistical noise, no problem, we'll get'em next time"
I stopped caring because seriously, we are talking about %0.0x changes overall. Notify me when changes are %0.1x
Quoting: EikeNobody's installing Windows on the Steam Deck.This is important, though.
Quoting: mr-victoryI stopped caring because seriously, we are talking about %0.0x changes overall. Notify me when changes are %0.1xMarch-April was a 0.14 rise. But, yeah, in general that's what the trend line is for: to see the signal amongst the noise.
Quoting: CatKillerQuoting: no_information_hereAs I have been complaining forever, Valve could easily take the data from live Steam client connections. They really have daily information of the base OS for every single user. It wouldn't have the hardware detail, but the OS statistics would be a 100% sample of logged-in users. Not sure why they don't.Valve know what OS each customer is using. Sharing that information with third parties without specific consent is a privacy violation.
I don't think sharing something like "2,295,123 of our 165 million users are using Linux" could trigger any privacy concern.
Quoting: no_information_hereQuoting: ArtenValve dont have actual data about all, because every month only subset of their user is in survey.
As I have been complaining forever, Valve could easily take the data from live Steam client connections. They really have daily information of the base OS for every single user. It wouldn't have the hardware detail, but the OS statistics would be a 100% sample of logged-in users. Not sure why they don't.
It's hurts more then help. Automatic telemetry shared with 3. party is no go in my case and probably it is relatively common opinion in linux community - more then in windows community. So it can cause bigger falsely underreporting of linux. One in year with exact data is best case and tolerable for me.
Quoting: ArtenQuoting: TheSHEEEPWith these, I always wish Valve would give out absolute numbers instead of percentages.
It's a bit like that "PC gaming is dead! Mobile is the future" nonsense that crops up every now and again - with PC gaming being in better shape than ever and with more games than ever.
Just because some share shrinks (or grows), doesn't mean there are actually less (or more) absolute users.
You only get this "very rough ballpark" number when Valve shares some actual user numbers once in a blue moon.
I don't think it is good idea. Valve dont have actual data about all, because every month only subset of their user is in survey. So % is better solution, because if they write absolute number, you expect real value.
Side note: -0.04% is so small number that can be be statistical error and reality can be positive change just because some bias in selection process or small sample size...
i think its more likely that the margin for error eat a bunch of windows users, then we got 0,04 as result.
eg, if the margin for errors was exactly 1% the true number is 1,04% not 0,04%.
there are a lot of views on how to install windows on the deck with plenty views
Quoting: gradyvuckovicHere's an alternative way of looking at the numbers:I think this is even an important way to look at it, as according to Valve users are counted only once per year.
September 2019 - 0.83%
September 2020 - 0.94% (+0.11%)
September 2021 - 1.05% (+0.11%)
September 2022 - 1.23% (+0.18%)
I don't know how they count the Steam Decks into this, my turn each year is August and though I had my Steam Deck by then, I didn't see a popup, neither in gaming mode nor in desktop mode. I guess they do it under the hood and you can't opt out?
Quoting: CyborgZetaInteresting that Ubuntu's back in 2nd place, and both Ubuntu 20.04 and Linux Mint have disappeared entirely.Not a surprise to me. Ubuntu remains the most popular distribution.
See more from me