Steam On Linux Drops Below 2% For August 2024 Survey

[phoronix]

Phoronix: Steam On Linux Drops Below 2% For August 2024 Survey

With the start of the new month comes the Steam Survey results for the month prior. The August 2024 data is in and it points to the Steam on Linux statistics dipping back below 2%...

https://www.phoronix.com/news/Steam-Survey-August-2024

[DumbFsck]

Inb4 hel88 comes and uses this noisy fluctuation to "prove" how much better Windows is or whatever.

All the while in the previous report (highest ever, iirc) we here were all talking about how much it fluctuates, etc.

[HEL88]

Inb4 hel88 comes and uses this noisy fluctuation

Let me just add that this fluctuation continues for the third month:
June: -0.24%
July: 0.00%
August: -0.16%
---------------------------------------------------------------

Getting back to my main topic - an addition to the survey:

Linux has: 1.92%
AMD AMD Custom GPU 0405: 27.79%
AMD Radeon Graphics (RADV VANGOGH): 12.79%


So GNU/Linux without Deck for the month of August has: 1.92 * (1 - (0.278 + 0.128)) = 1.14%

I wonder how this statistic would look like if W11 had a better reputation .

[ojsl1]

Valve really needs to be more transparent on what subset of users they hold these surveys for. I havent gotten a single survey prompt in my 5+ years on desktop Linux... Do they scan all users hardware nowadays?

[TheMightyBuzzard]

Does this actually matter to anyone besides the folks who've been screaming "Year of the Linux Desktop" for the past thirty years and change? Well, maybe the few private Valve investors.

[sophisticles]

So what we can conclude is that Windows is the OS of choice for the Chinese market?

[Etherman]

Valve really needs to be more transparent on what subset of users they hold these surveys for. I havent gotten a single survey prompt in my 5+ years on desktop Linux... Do they scan all users hardware nowadays?

I've got two surveys this year.
I run Steam maybe one hour or two hours a week.

[highball]

Valve really needs to be more transparent on what subset of users they hold these surveys for. I havent gotten a single survey prompt in my 5+ years on desktop Linux... Do they scan all users hardware nowadays?

Yeah, the survey is random per machine. You could have ten machines and 1 of them may never get the survey. You only really care about Year Over Year growth anyways. That's what indicates acceleration in growth. But Valve for sure knows what OS you are playing your games on. When you buy a game, the game is marked as a purchase of the OS type you were on to buy the game. After that, they check the hours you play the game and note how many hours you play on what OS type. The game is then counted as a game purchase for the OS type which has the most hours in that first two weeks. Also, a new feature of the Steam client in the review section can tell you if the reviewer was playing the game on a Steam Deck and how many hours were played on the Steam Deck. Valve for sure knows the exact number of Linux gamers and how many Linux machines with steam installed.

[pWe00Iri3e7Z9lHOX2Qx]

Valve really needs to be more transparent on what subset of users they hold these surveys for. I havent gotten a single survey prompt in my 5+ years on desktop Linux... Do they scan all users hardware nowadays?

I've gotten multiple surveys on different Linux machines this year. It still doesn't seem to capture the dGPU correctly in many cases. Both a laptop and desktop reported the iGPU and ignored the dedicated one.

The numbers are still interesting because there's basically fuck all in terms of useful telemetry for desktop Linux. So while not ideal, it's better than nothing to identify some broader trends.

[Mani]

Michael at least to me a graph showing the monthly data over the last two to three years (with maybe a running everage or trend line) would be far more interesting. Could you consider including something like that?