Linux market-share on Steam dropped 0.08% in December 2016

Nop, valve is faking this stats to don't scare Microsoft, and think linux is not a real threat. Meanwhile vale is developing a new steam machine generation that will break the market!

ok.. i have a big imagination.

@Jaromir: Linux user does not mean Linux gamer. Most Windows users are not gamers either.

Nop, valve is faking this stats to don't scare Microsoft, and think linux is not a real threat.

Very possible. The Steam "survey" is obviously not equal in means of count it pops up on Linux and Windows. Why don't Valve just use the everytime login information?

And Steam has definitely remained the natural place to go for Linux gamers.

Objection! I did not ditch Windows in favor of Linux to sell myself right to the "Microsoft #2" AKA "Valve". Will not touch anything DRM-related ever.

Linux dropped by 0.08% according to the figures.

It is a really interesting question... We are always talking about how popular is Linux between Steam users. Maybe it is a good idea to scoop how popular is Steam between Linux users?

I'm afraid such users really will be in the minority.

A question about this in the local survey should be interesting. Let's find it out!

And Steam has definitely remained the natural place to go for Linux gamers.

Not necessarily. Linux gamers tend to be averse to DRM and can avoid Steam altogether. The survey simply doesn't represent such users at all.

I'm afraid such users really will be in the minority.

May be, but a significant one I suspect, to make survey numbers not as valuable, as they are often presented.

Sure, I know Steam has itself been growing--but why would it be growing disproportionately among non-Linux users?

Why is this hard to believe? Linux gaming is growing, but there's no good reason to expect it would grow at exactly the same rate worldwide as PC gaming in general. And of course the survey won't give us exact figures. Trends are the most we can expect.

At exactly the same rate? No, to the contrary, there is reason to expect it would grow faster than PC gaming in general. Going back to the first point, it seems that as a proportion of total personal computer use (for browsing the web), Linux use is growing, according to a variety of stat-gathering outfits--the rate of growth reported varies, but the existence of growth seems to be widely agreed on. So the question is, would we expect the proportion of Linux users who game (on Steam) to grow faster than the proportion of Windows users who game, at the same rate, or slower? Well, I'm unaware of anything about Windows that has drastically improved the gaming experience in recent months; anything that might push console gamers or non-gamers to desktop computer gaming and Steam that I can imagine should work on both Windows users and Linux users. Anything that might push purchasers of games on CD to Steam should also work both on Windows and Linux users. Overall, I can't think of any source of overall Steam growth that should systematically exclude Linux desktops as opposed to Windows or Mac ones. And, again, the proportion of total desktops on Linux seems to be growing.

But I think we're all agreed that gaming on Linux has of late years and months been rapidly getting better; as a general desktop, Linux was already pretty mature a couple of years ago when it debuted on Steam, but as a viable gaming platform it has gotten way better in that time. Furthermore, distributions have now had a couple of years to get used to the idea that Steam is a thing, so they are much more likely to have it pre-installed or readily installable in whatever software repository they use than at the beginning. As Linux users update, Steam should have become steadily more accessible to them. So there are good reasons for a Linux desktop user to be much more likely to game, on Steam, using that Linux desktop now than a couple of years ago (and for dual-booters to spend more of their gaming time in Linux, less in Windows).

To sum up: If the overall percentage of Linux desktops is growing, and the viability of Linux for gaming is growing fast, then why would the percentage of Linux gamers be shrinking?

And Steam has definitely remained the natural place to go for Linux gamers.

Not necessarily. Linux gamers tend to be averse to DRM and can avoid Steam altogether. The survey simply doesn't represent such users at all.

I oversimplified, admittedly. My post was getting long. Yes, there are plenty of people who avoid DRM by using GOG and so forth. And my hat is off to them; I respect that. But they do seem to be in the minority, first of all. And more importantly, there does not seem to have been a sudden renaissance in purist principle among Linux users in the last year or so, so I would expect the principled Steam avoiders to stay in roughly the same proportion among Linux gamers over time. If anything I might expect new Linux adopters (of which there do seem to be a fair number) to be less likely to hold these principles strongly, and so the proportion to be, unfortunately, falling. So, GOG users might explain an oddly low Linux percentage on Steam in the first place, but not a decline over time.

To sum up: If the overall percentage of Linux desktops is growing, and the viability of Linux for gaming is growing fast, then why would the percentage of Linux gamers be shrinking?

I don't see why we should expect a direct, immediate cause-and-effect relationship. The rapidly improving quality of the gaming experience might be laying a foundation for growth, but we probably shouldn't expect an immediate surge in uptake.

To sum up: If the overall percentage of Linux desktops is growing, and the viability of Linux for gaming is growing fast, then why would the percentage of Linux gamers be shrinking?

I don't see why we should expect a direct, immediate cause-and-effect relationship. The rapidly improving quality of the gaming experience might be laying a foundation for growth, but we probably shouldn't expect an immediate surge in uptake.

One might say we shouldn't be sure of such a surge, but there's no reason to have no expectations of one. But stipulating no such surge, why should we expect a decline? What factor is it that is so not equal between Windows PCs and Linux PCs in the last year or so that although the percentage of PC desktops using Linux is apparently growing significantly, the percentage of PC desktops gaming on Linux (at least, on Steam) is supposed to be declining?
In all the times this stuff has been discussed here at GamingonLinux, I've seen exactly one factor proposed as an actual explanation: The proposal early in this thread that much of the recent Steam growth has been of Asian use of Steam, and that Asians don't use Linux. I don't actually know whether either of these things is true or, if true, a big enough factor to be an explanation.

But the claim I often see, that the Linux share on Steam is shrinking because Steam use overall is growing, is not an explanation. It presumes that some factor is causing Windows PC use of Steam to grow, that somehow does not cause Linux PC use of Steam to grow. But what factor? With the exception of the Asian idea I mention above which I am not at all sure I buy, nobody has suggested what such a factor could be.
Either there is such a factor, or there is something seriously wrong with the Steam survey. Since we don't know how the Steam survey works and we have lots of anecdotal indications that it is half-assed in various ways, I don't think we can rule out of bounds that second possibility. Alternatively, we need some suggestions about this mystery factor, or perhaps to shore up the Asian hypothesis with some figures. Because really, the idea that Linux Steam use (per number of Linux desktops) is shrinking fast compared to Windows Steam use (per number of Windows desktops) strikes me as very strange and counterintuitive.

It presumes that some factor is causing Windows PC use of Steam to grow, that somehow does not cause Linux PC use of Steam to grow. But what factor?

First of all, I don't see the survey as even representing Steam picture correctly (why do you need the survey, if Steam has complete data no doubt?). But one factor can potentially be higher number of games that don't have Linux versions still. If we assume gaming market is growing, and only a minority of games have Linux releases, then may be non Linux segment has larger growth?

One might say we shouldn't be sure of such a surge, but there's no reason to have no expectations of one. But stipulating no such surge, why should we expect a decline?

What decline? We're going in circles, but I still don't see why you'd expect every metric to grow linearly and in proportion to each other. There's just no good reason for this expectation.

I am sure, or rather I know Linux gaming is growing. Steam user base is also growing, and fast. These two might grow at different rates and those rates are not likely to be constant. Even in relation to each other.

Linux user base in general is growing.That's for sure.But every single person in this base are not gaming.Even some of them don't know Steam.

Linux usage is growing in great pace,also gaming on Linux is growing.But not that fast compared Linux user base in total.

I'd say gaming contributes to Linux usage growth in a big way, so relative percentage of Linux gamers is probably higher than relative percentage of Windows gamers.


Last edited by Shmerl on 16 January 2017 at 10:28 pm UTC

One might say we shouldn't be sure of such a surge, but there's no reason to have no expectations of one. But stipulating no such surge, why should we expect a decline?

What decline? We're going in circles, but I still don't see why you'd expect every metric to grow linearly and in proportion to each other. There's just no good reason for this expectation.

I am sure, or rather I know Linux gaming is growing. Steam user base is also growing, and fast. These two might grow at different rates and those rates are not likely to be constant. Even in relation to each other.

We are indeed going in circles.
Why would they grow at different rates? Let's start from the simplest case: The PC population stays constant, the percentage of that population running Linux stays constant, but the percentage of that population using Steam grows. If the general role filled by those PCs for Linux and Windows is either the same, or different but in a way that does not itself shift as Steam use grows, one would expect Steam's expansion into Windows PCs and Linux PCs to happen at the same rate, and so if the Windows Steam population doubled, so would the Linux population.

Factors could exist that would make that not be the case. They would have to involve some change in the roles of either Windows PCs, Linux PCs or both during the period of Steam expansion.

First of all, I don't see the survey as even representing Steam picture correctly (why do you need the survey, if Steam has complete data no doubt?). But one factor can potentially be higher number of games that don't have Linux versions still. If we assume gaming market is growing, and only a minority of games have Linux releases, then may be non Linux segment has larger growth?

What? No. That doesn't make sense at all. Linux already had that problem at the beginning of the market growth, so it was baked into Linux's earlier numbers--in fact the problem was worse before. So that should either have zero impact or work in the other direction. Honestly, I'm tearing my hair out here; everyone's giving me explanations for why Linux might have small numbers at a snapshot in time, and telling me they're explanations for change over time. Those things are different!

...one would expect Steam's expansion into Windows PCs and Linux PCs to happen at the same rate, and so if the Windows Steam population doubled, so would the Linux population.

This expectation is what I think is unfounded. There's just no basis for it. It's entirely possible that the vastly larger Windows population is currently taking up Steam more rapidly than Linux users. Or maybe the disproportionate change only happened during the Winter Sale.

Honestly, I'm tearing my hair out here; everyone's giving me explanations for why Linux might have small numbers at a snapshot in time, and telling me they're explanations for change over time. Those things are different!

Please don't tear out your hair. This isn't such an important issue that we need to get emotionally invested. We'll see in time anyway.

We should keed buying and playing linux games from steam... that will make the difference in the future.

...one would expect Steam's expansion into Windows PCs and Linux PCs to happen at the same rate, and so if the Windows Steam population doubled, so would the Linux population.

This expectation is what I think is unfounded. There's just no basis for it. It's entirely possible that the vastly larger Windows population is currently taking up Steam more rapidly than Linux users. Or maybe the disproportionate change only happened during the Winter Sale.

I'm not saying it's impossible for the Windows population to be taking up Steam more rapidly than Linux users (although the size of their population is completely irrelevant and a red herring).
But there is a simple, direct basis for the expectation of those rates of growth being similar: A desktop is a desktop and the population of users is the population of users. They are all part of the same market, they see the same ads or whatever--why should they be behaving differently?
Even if there is some systematic difference between the gaming tendencies of Linux users and Windows users, why should that difference itself be changing? Imagine Linux desktop users have a lower tendency to game, such that half as many Linux users per 10,000 desktops play games. Then, if Linux had 2% of the desktop market, it would not be unreasonable for Linux to be at 1% on Steam. But then, if Linux increased to 3% of the desktop market, one would expect Linux to rise to 1.5% on Steam.
Instead, we have apparently a long-term trend of in the neighbourhood of from 2% to 3% on the desktop (ish, vaguely--whatever the specifics, like maybe 1 and a half times as much as a couple years ago) and a simultaneous long-term trend of from 1% on the Steam survey some time back, down to about 0.8% now. Monthly shifts have generally been insignificant, but the long term trend has been downwards. For these two things to both be true, Linux users would have to be gaming at lower and lower rates compared to Windows users over this time period, or at least adopting Steam at lower and lower rates. Or all of Steam's growth would have to be coming from some particular population which includes a much lower percentage of Linux users.

But we're talking about a big shift here if we accept Steam's figures, of Linux users gaming on Steam a lot less, like half as much per capita, from Linux desktops being used for gaming maybe half as much as Windows ones (1% Steam/2% desktop), to Linux desktops being used for gaming about 1/4 as much as Windows ones (0.8% Steam/3.?% desktop), over the course of a couple of years. Would you not expect such a trend to be visible in some way? Is it not reasonable to expect some explanation for the shift to exist, especially at a time when all our information suggests Linux becoming far more viable for gaming, rather than less?

Or alternatively, one could conclude the Steam figures are bogus in some way. I've reserved judgement for a long time, but I am gradually coming to the conclusion that that is the simpler and less unlikely explanation.


Last edited by Purple Library Guy on 17 January 2017 at 6:09 pm UTC

@Purple Library Guy: Yes I know you expect Linux users to be taking up gaming and Steam at roughly the same rate that Windows users do. I do get your point. I don't agree on your conclusions, but I do understand what you mean. The steam figures could be bogus, or they could be fairly accurate. We haven't got the data to make that call. All we can do is speculate.

I stumbled upon this article about the growth on Steam, the data coming from the Steam Dev Days 2016 (didn't bother looking for an official URL for the slides):

http://boilingsteam.com/steam-dev-days-organic-growth-and-rising-asia/

The one feasible solution to this problem is regional growth differences. Assuming that the dominating markets w.r.t. growth have lower Linux market share, it is obvious that Linux shares on Steam drop overall.

We could now start to discuss which regions have more or less Linux market shares, but I don't know of anything beyond prejudice to back any claims.