Valve Now Reports The Steam Linux Marketshare At Just 0.35%

[stalkerg]

Ohhh I know now:

Simplified Chinese
56.37% +26.83%

English
21.24% -13.40%

Also, AMD drop from 19% to 10% (on Windows)

[schmidtbag]

This is very inaccurate. None of my 3 linux boxes have EVER been part of the survey.

What they need to do is get anonymized data automatically, so we can finally see accurate numbers.

Around 2 years ago, I got the survey multiple times per year. Now, I think I only got it once in the past year. Definitely something weird going on, though, I have also been signing on less often.

I don't think they need anonymity, but rather have a checkbox stating you agree to have your hardware automatically surveyed. That keeps everyone happy - both the tin-foil hat wearers and those who care about the data. Then, the first time you sign on for the month, your computer sends your hardware data. The only people who are left out are those who intentionally opt-out (who would be a minority) and those who didn't sign on for the month, who are irrelevant anyway.

Unfortunately, Wine users will always remain being identified as Windows users.

[Luke_Wolf]

Something wrong:
Windows 7 64 bit
+22.59%

Windows 10 64 bit
-17.14%

I can't like this post twice so I'm also going to quote it. This so so very much.


Michael you can't claim that this survey is the end all, be all. Sorry but no. It's well known this survey has major methodology issues which result in things just like this, and this is very far from the first time we've seen such irregular results. If Valve wanted to have actually valid results they could easily do a population survey rather than just whatever the heck kind of sampling they're doing, which nobody has been able to figure out yet, and they won't talk about.

[stalkerg]

Looks like you got it right. Look at the language section. Chinese +26.83%

I think no, because this version is Windows10 but rose only Windows7.

[stalkerg]

Perhaps some good games pushed to the China market in Steam?

[ThoreauHD]

I literally just downloaded 240 GB worth of Linux steam games not 30 minutes ago. I don't know where they're getting their stats from, but reddit is flooded with windows and apple users moving to steam on linux. I wish that .04% or whatever the F they think it is was reflected in the ridiculous amount of tech support questions they keep posting. It's fking exhausting.

[boxie]

As any statistic sampling goes, it is by definition wrong, but still good enough to get a general idea.

Also, the steam client has a browser embedded in it, I wonder if its browser strings are also collected/used to determine OS usage?

[ramrod]

I think no, because this version is Windows10 but rose only Windows7.

I wonder where the sudden influx of Chinese users come from then. I doubt bots would bother with a survey, so it's most likely real users.

[humbug]

I wonder where the sudden influx of Chinese users come from. It's an increase of close to 3 million users if my math is right. I doubt bots would bother with a survey, so it's most likely real users.

http://blog.counter-strike.net/index.php/2017/09/19277/

[DarkFoss]

I've purposely not booted into Windows at anytime during the 1st of the month since April . I do several re-boots of Linux on those mornings multiple re-starts of Steam, bought a few Linux games ect . Nothing triggered a survey . The only time I've had one is when I had to install Ubuntu/Steam 2 days in a row mid September.

Perhaps it's time to have an install day via vm/flatpack keep doing it until a survey is triggered ? Try and generate a true reflection instead of grousing about the inaccuracies each month after these numbers come out.