Steam has finally published its latest findings across its gargantuan platform, and as you’d expect there’s some interesting tidbits about millions upon millions of gamers who use the service.
Windows 10 has finally reached the tipping point
It took a while, but Windows 10 has finally reached the point of no return. The 64-bit version of the OS was used by 50.15% of respondents, which is nearly twice the amount of users that have refused to move off Windows 7. Interestingly, despite being universally admonished as total garbage and getting bugger all support from developers, 0.1% of users surveyed were still using Windows Vista. Sounds like some cheeky office workers sneaking Nidhogg and Spelunky onto their office PCs, if you ask me.
The adoption rate of VR … isn’t much
Both the HTC Vive and Oculus Rift became a little more popular from February to March, with little being the operative word. The Rift’s userbase jumped by 0.02% among respondents to 0.13%, while the Vive remained top dog with 0.24% (up by 0.01%) of respondents owning the Newell-sanctioned headset. The survey didn’t outline any other competitors bar the devkit variations of the Rift, however, although it’s difficult to imagine the userbase of headsets like StarVR, OSVR and LG’s prototype challenging the Rift or Vive anytime soon.
NVIDIA’s GTX 970 is still the most popular gamer GPU
Whether it’s a system playing predominately DirectX 10, 11 or 12 games, the end result is the same: NVIDIA’s GeForce GTX 970 is the most popular GPU among enthusiast gamers. It’s not the most common GPU across the whole platform – that’s the HD 4000, the integrated GPU that shipped with Intel’s Ivy Bridge CPUs around 2012 and early 2013. But for gamers who want something a bit more, the GTX 970 is the way to go. It’s the top GPU amongst DX10, DX11 and DX12 systems, with the affordable GTX 960 runner up for older PCs.
For PCs on Windows 10 playing DX12 games, the GTX 1060 and GTX 1070 have become the second and third most popular GPUs. The results are a good snapshot of what PCs use the further you go down the list: there are more people, for instance, running the mobile variant of the GTX 960 than the high-end GTX 1080. The GTX 970 is also the most popular GPU among Linux users, beating out the HD 4000 and HD 4400.
As for AMD’s Polaris GPUs? The RX 480 is only used by 0.70% of respondents on Windows 10 PCs, and 0.81% of respondents on all variants on Windows. The real interesting part here is where the HD 7700, HD 7900 and HD 8800 GPUs, cards that are five and four years old respectively, was more commonplace in the survey than last year’s RX 480. For PC gamers, that’s a long time between upgrades.
Intel is getting more popular – for now
Nearly 80% of respondents were using an Intel CPU, a 1.46% bump from the previous year. That’s a substantial chunk of the market, but it’s also worth considering that the AMD Ryzen CPUs only launched in market at the start of the month. Most importantly, the Ryzen 3 and 5 series CPUs aren’t available yet. And those are the CPUs targeted at the biggest chunk of gamers on Steam: just over half of respondents were using a system with only 4 CPUs, with almost 44% on a system with only 2 CPUs.
Given that the Ryzen 1700, 1700X and 1800X are targeted at developers, streamers, desktop workstations and gamers who dabble in between all three, it wasn’t likely that Steam’s survey would reflect any kind of shift towards AMD in the March results. (Only 0.32% of users surveyed was using a hexacore PC, which gives you an indication of what the PC market really looks like.) We’ll get a better idea of how the Ryzen CPUs are received across Steam towards the end of the year, when all of the Ryzen CPUs have begun to disseminate throughout the market.
Nearly 18% of people don’t know whether they have a microphone
This one says more about the people using their system than the survey, I think. Needless to say, 17.93% of those surveyed weren’t sure about whether they could spout profanities to their fellow gamers on Steam. 18.67% of people knew for sure they didn’t have a microphone at all – or just preferred to say they didn’t.
Nearly 1% of Steam users are still using Windows XP
God bless your souls, you fine people.
Comments
35 responses to “Valve’s Steam Survey: Windows 10 Is Popular, VR Is Not”
970 is a pretty solid card, I am completely unsurprised.
My little 670 still troopering on!
I am the 0.41%
Filthy sub-human, six series user. 😛
660 here. Not even sure if it’s the Ti version at the moment.
Wow, twice the amount of people are using XP than the Vive and the Rift combined. Let’s hope that tech can be saved by other industries.
Let’s hope it dies along with 3D Movies.
Virtual Reality (how weird is it to actually type that out, the term has been overtaken by its own abbreviation) has been riding a wave of good PR, for a long time. It’s just not cutting through.
The peripheral (no pun intended…… 😀 ) distractions like Luckey/etc should be largely ignored in favour of decent software to take advantage of the amount of chatter in the games press/streaming communties surrounding it.
But that’s not happening. Instead of continually trying to charm us with VR, I hope we get to see more people talk about why it continues to fail so spectacularly.
For years, publishers and the press were falling over themselves to build up the ‘identity’ aspect of the console/pc industry and foster communities so we could all band together and play everything that had multiplayer bolted on, and then we got system-level tat like achievements and trophies to show off to all our paying subscriber mates.
Now, they’ve gone the other direction:
You. All alone. Wearing that headset. Emotional. Personal. Solitary.
I don’t know the exact name for this weird about-face, but as someone in the audience I think it’s called whiplash.
Lost opportunity to say “VR has had good PR”. 50 DKP minus.
It didnt help that the community that has fractured around the two main headset are acting like a bunch of monkeys throwing poo at each other. Not only does it make the technology look immature, but the community too.
But it is very much an enthusiast’s market right now. As long as companies keep putting investment into improving the tech, the rest will follow.
As as for being solitary, not so much. Apps like Bigscreen Beta allow virtual LANS. The more users out there, the more opportunity for sharing experiences.
I don’t see how there is any kind of about-face at all. Both involve sitting at home alone playing games in solitude – all the multiplayer that was being pushed was online, not splitscreen. In fact it’s one of the bigger problems facing VR games at the moment – there are way too many games that are multiplayer-focused (no AI = much easier to make), but the install base isn’t large enough to support so many different multiplayer games so it’s often pretty hard to find a match if you want to play something that isn’t the latest and greatest.
VR’s problem is that its still to expensive and the fact that its a Peripheral thats being treated as seperate consoles with exculsives because of 2 gigantic arsehole companies refuse to play nice
Only Occulus has engaged in anti-competitive behaviour such as proprietary APIs and bribing developers for hardware exclusivity.
I’d argue it’s only one gigantic arsehole company.
And Steam’s API is proprietary as well. It is not open source … it is open licence as well. Big difference. And does not support all functions of the Rift, such as ASW.
As for the “bribing developers” …. it’s called FUNDING DEVELOPMENT. It’s something that Valve doesn’t do. Developers have got to eat too, you know.
And if their funding development, there has to be a payoff. If that means timed exclusives, then rhats a small price to pay.
The implementation is proprietary, the API is open source under BSD. Anyone can write their own implementation if they want to.
And yet the core is tightly controlled by Valve. I’ll wait for OpenXR which is an implementation that is not controlled by a single entity.
https://uploadvr.com/serious-sam-oculus-offered-shitton-money-rift-exclusivity/
It’s not “funding development” when the game’s already been developed for both platforms by a reputable, stable developer. I understand their are serious costs in software development, and for many people everything is on the line getting the product out the door with no guarantee on return. That was not the case here.
It’s one thing for Occulus to pay for content to be created for their platform [and, sure, exclusivity is a reasonable assumption of said paid development]. It’s another to throw money at an already developed game for it to be made exclusive on their platform.
Regarding the API, I was going to link to the openVR github page, but @zombiejesus beat me to it.
If a company can afford to fund the software development on their own, then great! Croteam could afford it. Other developers can’t. As such, there is a lot more VR content out there now than there would be otherwise.
As I pointed out with the core of the API, Valve still controls it. OpenXR will at least be controlled by more than one entity …. and Valve and Oculus will have an equal stake.
That’s pretty much what I said anyway. What about when they throw money to developers who don’t need it to survive, just to gain exclusivity by throwing money at it. Buying exclusivity doesn’t help the ecosystem, it only helps to create a monopoly.
OpenXR will be great. But for the last 2 years or so [and for at least the next year until OpenXR 1.0 spec is released], Valve’s still the most open player out there.
So how exactly does Oculus know if a developer needs the money? One way is to offer it …. and if it’s rejected, they don’t. As for creating a monopoly, Steam’s already pretty close to one. You have to break into the market somehow.
Not really seeing the problem with offering exclusive deals. Developers have a natural bias against exclusives and will always prefer to sell their games to the most number of people, if they can, but if they’re short on cash then it’s good that they may also have the choice of pre-funded development available if they want it.
I’m surprised there are so many machines packing an HD4000. What are they playing? Is solitaire on steam?
It’s not 2004 any more. Gone are the days of the completely anaemic IGP which is only good for office applications.
The current generation are integrated into the GPU die, so they gain a lot of performance simply through close proximity to the CPU and memory busses. That said, they are actually quite capable for the vast majority of games on steam.
Sure, they won’t play the latest AAA title on ultra settings, but you will have a playable experience on plenty of other titles.
AMD’s APU’s are even more capable again.
But, going back to the HD4000. My main laptop has an Ivy-Bridge CPU [i5 5200u], which has two cores with hyper threading and is a low power [15W] unit. I run Ubuntu 14.04 with the standard intel open source drivers. All of Valve’s catalogue of games runs perfectly fine on medium-ish settings. CS:GO, Dota 2, portal etc are all perfectly playable.
Again, it’s not a powerhouse, but to mention Solitaire is a bit of an insult to what these great little chips are capable of 😉
Man you need that piece of hardware to play Minesweeper at 60FPS in 640×460.
They surely play Dota, WOW 3 and like. Because they are playable. Probably AgeofEmpires, Counterstrike, you would be surprised of what is playable. Sports: FIFA 12,13, Civilization. Chess,WOrms. Probably not a lot of solitaire.
Source some others:
https://www.intel.de/content/www/de/de/support/articles/000005659/graphics-drivers/graphics-for-3rd-generation-intel-processors.html
Lies, damn lies, and statistics….. You can make many an argument based on interpretation of stats.
It’s pretty easy to paint a picture from this article that the average “enthusiast” gamer is sporting an Ivy Bridge i5/7 + 970, a pretty powerful PC when new, all the while being welcoming of the move to windows 10.
Looking a bit further into the details, a different picture can be seen.
The most standout for me is desktop resolution. Sure, most popular is 1920×1080 at 45.26%, but in second place is 1366 x 768 at 23.26%. That screams entry-level laptop for sure.
This also ties into the top GPU stats. The 970 may be the most popular external, discrete GPU, but tying it to much else is a bit of a stretch.
The top 4 most popular GPUs are all intel integrated GPUs, and account for 50% of the total. Now, that may be skewed a little if it counts both an IGP as well as a discrete, but I’m pretty sure the survey only takes the one being used for the primary display. The most popular VRAM amount is 1GB!
Either way, seems to me the average gamer runs an entry-level laptop they have for school or work, with an integrated GPU and most likely shipping with windows 10. That’s more than enough to play most games at decent speeds [particularly at 1366 x 768 resolution]. Dota 2, CSGO and the like will run more than well on such systems.
It’s a long way from someone running a 970.
You probably don’t want the average on bracketed data like this, the median is more useful. For screen resolution the median is in the 1920×1080 bracket (only 48.14% of users have resolutions below that) so the median gamer is running 1920×1080. The reason you probably don’t want the average is because the average screen resolution is 1703×980 (which is closer to 1920×1080 than it is to 1366×768).
You can’t average the GPU brackets at all unless you assign them numeric scores somehow, but you can find the median (which I’m not going to do because there are way too many of them to bother). We can kinda see whether the median is in a mobile or desktop GPU though – 15.67% of DX12 GPUs are mobile, and DX12 GPUs make up 74.76% of all GPUs. To be able to hit the median mobile would need another 38.2% of total share but only 25.24% of GPUs are non-DX12 – in short, it’s guaranteed that a desktop GPU is the median GPU, I’m just not sure which one specifically.
The term “average” doesn’t really work here I suppose.
A more accurate term would be the “majority” of users are using integrated graphics on a laptop.
I understand what you mean, I did outline why the data doesn’t support that though. Out of all GPUs surveyed, 63% are desktop DX12 GPUs which puts laptop GPUs in the minority. The 1366×768 resolution is an interesting one but it turns out there are quite a lot of desktop monitors that have that native resolution. They’re dirt cheap and I’d never buy one myself, but it could account for some of that 23.26% resolution share.
Oh yeah I see what you mean. I didn’t realise that the percentages under each GPU type were only for that particular category [ie: DX11 or DX12], not a percentage of total.
It’s still hard to work out which are desktop and laptop, though [apart from the “M” designation in nVidia]. You can get non-‘M’ model GPUs in laptops, and AMD has no designation other than the series.
The desktop resolution still speaks volumes though. I know you stated there’s lots of cheap desktop monitors that are 1366×768. Perhaps, but there’s no way anyone would get one unless they didn’t know any better [ie: picking up a pre-built from Harvey Norman]. And even then, you wouldn’t have a 970 in a system with that monitor [I would hope?].
I still pertain that 970 [and equivalent high-end GPU/CPU combo] based systems are still in the minority. And by no means represent a snapshot of the typical steam gamer.
But hey, who knows, right? That’s pretty much what I was really saying I guess, any conclusions derived from statistics always has some sort of assumptions thrown in.
I did my best to separate out mobile and desktop GPUs but AMD doesn’t make it easy. Especially since with some models they use the exact same model number, which just doesn’t help anyone. I don’t think any missed ones would be enough to push mobile over 50% though.
I agree that some conclusions need to be educated guesses or ‘most likely’ situations, but it is possible to get some concrete data out of the stats. Plus they’re just generally quite interesting, like seeing how badly AMD is doing in the gamer market against Intel and Nvidia.
Yeah, the CPU stats for intel vs AMD seem about what I’d expect. 80/20 split, seems about right given the overwhelmingly poor performance of Bulldozer onwards.
Will be interested to see if Zen brings some people back to the AMD camp.
Once again I think the data may obscure things in the desktop GPU scene. I have no doubt nvidia would be in front, but I’d be incredibly surprised if it was anywhere near 80/20. I’d think more towards 60/40 range or better. The laptop graphics most likely skews this out quite a bit though. It’s pretty tough to find AMD powered laptops [which is a shame, their APU’s are actually nice little low-power units].
On that note, why are there no nVidia GPUs in the DX11 section? [Are all of their DX11 GPUs capable of DX12 as well?].
They made DX11 cards software-upgradeable to support DX12, there weren’t any hardware limitations that would prevent it. Nvidia has a handful of cards that are stuck at DX11, I’m more surprised how many AMD cards weren’t updated. Hardware limit on the AMD design maybe?
So we’ve reached the comment depth limit of kotaku, haha 😉
They were updated in the exact same way nVidia’s were.
AMD’s 6900 series and below use the TerraScale microarchitecture, whereas 7000 series and later moved to the GCN microarch, which is radically different.
The 7000 series was released in 2012, initially as a DX11 card. All subsequent cards have received DX12 support through driver updates.
Given the relatively tiny number of DX12 games, I’d say it’s clearly not worth the time to support it on such old GPUs.
‘Free’ Windows 10 upgrade vs a couple grand for the VR experience i wonder why 1 is more popular then the other
I’m pretty damn sure, if valve actually would make the vive cheaper and actually make games for the thing, VR would go far. Been waiting a while now for a decent game, can’t keep playing Arizona Sunshine because its the only good full game out
I think the Rift got a $100 price slash recently so pretty sure Vive will become cheaper. Not that I care for 1200p gaming, I have lived in the world of 4k for too long and now see the voids between pixels quite clearly on lower resolution screens. (unless its a real small screen, but I don’t accept anything below 40″ these days.)