Steam Survey for February 2024 shows a big bump in Simplified Chinese

You'd think, after the second, third or possibly tenth time it happened, someone at Valve would look into the why of it and you know, fix it.

When will Valve start selling the Steam Deck in China? It's strange to ignore such a big market.

Quoting: sarmadWhen will Valve start selling the Steam Deck in China? It's strange to ignore such a big market.

I said in a comment on a different article

Quoting: CatKillerI think having an official third party releasing in Hong Kong is as close as Valve wants to get to an official release in China. They've had to make a Steam For China since Chinese customers were using VPNs and vanilla Steam to get round the restrictions on gaming in China; there are 30-something games released per day on Steam, but around that many are authorised for sale in China per year; if you're too young it's a legal requirement that you're only allowed to spend a specified number of hours per week gaming, and only at specified times. It's just way too many headaches all round to try to sell a vanilla-Steam-based gaming machine whose main attraction is that you can play on it whenever you want.

That's still pretty much what I think.

I would really like to see Valve work out how to ship the Deck to more countries - the fact that people that want one can't get one is the biggest flaw of the Deck. Whether that's with a logistics partner like they've got in those Asian territories, or doing it mostly themselves like they are in North America and the EU. But China, as big a market as it is, comes with its own special headaches.

Does anybody have detailed Steam survey data stored somewhere? I wanted to look into the percentages when filtering by Linux only, and the Internet Archive's data is sparse on this one...


Last edited by Eike on 5 March 2024 at 11:57 am UTC

Last month, I've been engaged in an argument about how to properly show what's happening in regards of Linux adoption, and I want to present a different view on it that IMHO gives a better feeling for what is happening (and what is not happening).

I have been using the HoloOS percentage of the Linux-filtered results to calculate Steam Desk adoption. (*) I removed three spikes from the overall Linux values (you can see them in the article above), and I didn't have full information about the Linux only results, as "Internet Archive" doesn't have them all.

But here you go...



(*) The GPU might have been even more precise, but the difference is very little (second digit after comma when regarding only Linux values, so it's something like 22.77% vs 22.75%), and unfortunately, the GPU seem to have been renamed in the results at least twice.


Last edited by Eike on 5 March 2024 at 3:37 pm UTC

Quoting: EikeBut here you go...

That's not a bad approach. I've filled in the gaps for you.



(I'd previously discussed with Liam dropping the definitely-bad 03/23 from the tracker and he was insistent that the data we get is the data we get, so I've got the dips you removed in for consistency with the tracker data - same as I've done with the Linux/Mac graph previously)

Note that because this is a record of proportions, if the trend for desktop Linux were completely flat (it's not - it has a positive gradient) it wouldn't mean that desktop Linux usage wasn't growing, only that it's growing at the same rate as Steam as a whole. Since the trend is slightly positive, it means that desktop Linux usage is growing slightly faster than Steam as a whole.

Quoting: CatKillerThat's not a bad approach. I've filled in the gaps for you.

Cool!

Where did you get the missing Steam Deck percentages? Internet Archive didn't yield data for many months...

I guess you added the trend lines manually?

Quoting: CatKiller(I'd previously discussed with Liam dropping the definitely-bad 03/23 from the tracker and he was insistent that the data we get is the data we get, so I've got the dips you removed in for consistency with the tracker data - same as I've done with the Linux/Mac graph previously)

I wasn't sure about that either. On the one hand, it looks like bad data, and we know there's problems with the data, on the other hand, leaving something out feels like cheating. *shrug*

Quoting: CatKillerNote that because this is a record of proportions, if the trend for desktop Linux were completely flat (it's not - it has a positive gradient) it wouldn't mean that desktop Linux usage wasn't growing, only that it's growing at the same rate as Steam as a whole. Since the trend is slightly positive, it means that desktop Linux usage is growing slightly faster than Steam as a whole.

I'd say, percentage-wise nothing has happened for over two years, but as you're pointing out, Steam is probably growing, so yes, desktop Linux usage is at least growing along. Good point.

Quoting: EikeWhere did you get the missing Steam Deck percentages? Internet Archive didn't yield data for many months...

Having dredged this data from the Wayback Machine before, I can give you two top tips:
There are two valid URLs for the hardware survey - one with "Steam-Hardware-Software-Survey-Welcome-to-Steam" and one without. The Wayback Machine archives both of them at different times.
The survey lists the delta from the previous month, so if you've got the value for one month you also know the value for the previous month.

Quoting: EikeI guess you added the trend lines manually?

Well, it's just an "add trend line" in LibreOffice. But it's the least-squares line of best fit, which is what you'd want.

Quoting: EikeI wasn't sure about that either. On the one hand, it looks like bad data, and we know there's problems with the data, on the other hand, leaving something out feels like cheating. *shrug*

Yeah, either would have been valid choices, I think. I gave Liam enough of a push so that it was something he'd definitely considered and made a definite choice about, and then I've gone with what he picked since.

Valve used to correct the survey results later in the month when there'd been an overcount of China, but they seem to have stopped doing that for the past 12 months or so. That's why the later data is so much noisier than the earlier data.

Quoting: EikeI'd say, percentage-wise nothing has happened for over two years, but as you're pointing out, Steam is probably growing, so yes, desktop Linux usage is at least growing along. Good point.

Eyeballing the desktop trend line now and comparing it to how I recall the trend line was when I was encouraging Liam to put the trend line on the tracker graph, it's steeper than it was then - certainly no shallower. Which means the growth rate would be higher now than then. One could check the gradient of the line of best fit over time if they wanted, but "line goes up" is good enough for me right now.


Last edited by CatKiller on 6 March 2024 at 11:06 am UTC