Valve Now Reports The Steam Linux Marketshare At Just 0.35%

[Kasoroth]

I'm using Linux only and I'm gaming on a daily basis via Steam (mainly Dota 2) and I get the survey regulary...

I use Linux only (no Wine), and game regularly, but I've only gotten the survey twice ever, and both times were shortly after doing a new install. Once was soon after Steam was released for Linux, and the other was after a clean install when I replaced my hard drive with a solid state drive. I've never gotten the survey any other time. I'm kind of curious about how often some people get the survey, because I've hear other people also mention getting it "regularly". Is it every few months? A couple times per year? Once every couple years?

It seems odd to me that a random sampling would have that much variation where some people get it "regularly", and others almost never. It seems like over a span of years, a random sampling would tend to average out so that any particular person would have roughly the same average frequency. If I roll a fair die once, a 1 is as likely as a 4 (or any other number), but if I roll the same die 60 times, the average is very likely to be close to 3.5.

The same principle should apply to the survey sampling. In any given month, random chance might just not give you the survey, but over the course of 60 months, it should tend to average out. Sure, there will still be occasional outliers who got exceptionally many or few surveys, but the vast majority of people should be relatively clustered around an average number of surveys that is based on what percentage of the users they sample each month.

Bad sampling might be equally effecting all operating systems, or it might be biasing the results, but if we can't trust the sampling, the data is basically garbage, at least with regard to OS popularity.

Personally, I'd like to see them include a few basic things like OS and processor in an automatic full-population survey that's automatically on by default, with an option to opt-out if you want. I'd also like them to include an account setting for "preferred OS", to capture information about how many people are using an OS other than their preferred one, along with which specific games they play in their non-preferred OS. It seems like this could be potentially useful data to game developers. It might be interesting to measure what percentage of people's usage time is in their preferred vs. non-preferred OS.

It seems to me that the current information they're collecting is of limited use, possibly inaccurate, and potentially more misleading than illuminating

[Bonnom]

Way more chinese users and way more new Player Unkown Battleground users, a windows only game.
PUBG doesnt run very well on windows. It does use unreal engine 4 and could relatively easy be ported to linux.
But judging from how well the game is optimized for windows, I don't think the programmers are good enough to do a proper linux port.....

[OneBitUser]

You've got this backwards, the survey is used to determine the proportion of users on Windows (vs those on Mac vs those on Linux). If in the entire population of the set, in this case all Steam users, 98% are actually on Windows, you'd want your random sample to be 98% comprised of Windows users. This way your sample would be representative of the population, that's the aim, to establish percentages/ratios. So if the survey hits more Windows users, that makes it MORE accurate, because we can take it as given that amongst all Steam users, most are on Windows.

To be clear, I'm not arguing the Steam survey is accurate or the methodology is sound. I'm just saying that the supposedly counter-arguments woefully presented and make no sense if you possess school-level mathematics or higher.

I do not have it backwards, I assure you. I understand how and why the survey is distributed, i.e. at random to a number of people, until it gets a certain number of completed surveys. Theoretically the randomly sent out surveys should represent the entire population. Unless of course, your targeted number of completed surveys is a way too small portion of the total population. In which case, your results will most likely over represent the majority part of the population (look at the Win7 and Simplified Chinese numbers), because the randomness favors the majority. Until you send out a large enough number of surveys, that is.
The surveys should also take into account the number of active players in a given region in a given timeframe: if X% of players log in from Asian addresses (which percentage Valve should know without a survey), then only X% of the surveys should be completed by them, not more.
Steam has several hundred million users, with 10-16 million users online in peak times, and easily over 20 million people online daily. The surveys' sample is in the 5-digit range, if I remember correctly. That might not be enough to represent a population this large, and considering the fact that it is digital and highly automated, Valve could simply aim for a sample of say 5 million people, geographically distributed to represent the whole population.
I guess that would give a more accurate figure of 1-3% Steam Linux users.

[Dukenukemx]

I like how many people here are saying they never got the survey or something about Chinese players, but the truth is nobody wants to use an OS where they can't play ALL their games. Wolfenstein 2 is out and is it for Linux? It runs OpenGl and Vulkan right? And yet, not on Linux.

We need two things. For one, a way to play Windows only games on linux. Forget Wine, it sucks. Secondly, we need to force developers to make a Linux version or have Valve port the games themselves. It'll never take off unless it's super easy for users.

[vito]

This is a chicken and an egg problem - many companies do not publish games for Linux because the market share is small, and the Linux market share will be small until there are more games available for Linux.

The only way to change this is to vote with your wallet and buy only Linux games from companies like Valve and Feral. Once people stop buying Windows only titles, profit will go down and companies will come running. This is the only way. It is free market at work!


I find it amusing when people that keep playing on Windows complain that there are no Linux games. It reminds me of the people that keep buying Intel CPUs and keep complaining about their prices.

[carewolf]

"I've never had the survey so it must be flawed"

^ people who don't understand probability.

Users are surveyed on average 1-2 times a year. If you are not surveyed and have been on steam for more than a year, something is off. You would probably trigger a survey when reinstalling. I have been surveyed 4/5 times when installing steam on a new machine, and all the other times was with a steam upgrade. So if a large group of new users install steam they were all likely surveyed this month when they joined, where normally they would only get surveyed every 10 months.

It is pretty crap and random.

[OneBitUser]

The only way to change this is to vote with your wallet and buy only Linux games from companies like Valve and Feral. Once people stop buying Windows only titles, profit will go down and companies will come running. This is the only way. It is free market at work!

I have been shopping with that in mind for a year now, and even previously most of my purchases were for Linux titles from Linux machines.

But this is a typical case of the gullible, uninformed masses of consumers making the decision for everyone... even the well informed and conscious consumers. As long as the majority remains an ignorant consumer it does not matter what I, as a conscious and informed shopper do: the market will continue on its present course, simply because it can.
Stupid people are making the decision or everyone, because they remain ignorant of DRMs, DRMs within DRMs, compulsory GeForce Experience usage, and compulsory constant internet access, etc..

[Mez']

The flaw in your logic is that the english windows userbase hasnt shrunk at all. 50m english users out of 150m total is 33%... add in 50m new chinese accounts and now english is only 50m/200m, -> 25% a drop in 8%, and chinese went from almost 0 to 25%, a gain in 25%....

My numbers are purely made up just to demonstrate that you can have downward swings in percentages without losing any absolute value. This also obviously explains why linux numbers are tanking. Valve has said in the past that their chinese numbers are growing fast. The population of China is at 1.4billion... 50 million people is less than 5% of the population, about would be perfectly reasonable amount of spike if there was some discount somewhere, or an advertisement.

Turn it any way you want, such a big part of the chinese population didn't just show up on Steam all at once, out of the blue.
Even with such a large population, penetration of the market always takes a long time between early adopters and even early majority, a month is out of the question.
It doesn't mean Linux figures are wrong but such strong variances do not happen at this stage of the product life cycle (even if some local branches can still be on growth). econometrics-wise. The only way to get such variances is by sampling bias.
Which leads to questioning how their statistics are obtained since they don't make any sense econometrics and marketing-wise.

Look at the browser or general OS stats and curves, they've never been turned upside down in such a way from one month to the other, even when China went crazy on pirated old Windows versions. This is a steady progress that takes time.

[aht0]

Even on Windows I've had survey maybe twice, counting all the years I've had Steam account. Since CS:S release, whenever it was.