Latest Steam survey sees AMD and Windows 11 crash as a new top language appears

[LimyG]

P.S. I did actually participate of my own free will a few years ago. Just once.
It still causes me trauma, and waking up in a cold sweat. The most stupid thing I ever did!

 

[FaTaL]

Steam survey is *highly* suspicious. This month the percentage of Chinese speaking users went up by 13.71 percentage points. All of a sudden they added tens of millions of new users and only in China!? That happened a few months ago, with many inexplicable large jumps in various categories, only to be completely reverted a month later.

bots. I don't trust metadata anymore.

 

[LimyG]

bots. I don't trust metadata anymore.

Fully with you on that.
I don't mean to broaden the topic, but can't help myself. Who or what will we be about to trust (except for family and close friends) in 5 years time.
Will info be non-human - as in artificial. Sceptisism is a wise policy with anything on. Well it always has been to a degree. But Oh the innocence of the dial up 56 k/bs internet, real speed 40 k/bs was considered decent.
I guess some folks here don't remember that, but at that time viruses existed primarily in email. The internet had few pictures, was largly text, but how big is an average, if such a thing, net Malware.
It wouldn't have had time to down load.
Those who wern't on line at that time, just think, even a crappy sustained 1mb/s net connection is 40x faster.
Most people are at over 10 m/bs, 400x faster. It was a different world where a degree of trust existed on the net.
Now it's guards up. Not as bad as it sounds, most people know how to protect their computers, but very occaisionally a post like @FaTaL reminds me the difference.

Whatever, progress is normal, so I'll shut up about that until a year later or so.

 

[EndRessentiment]

But does it matter how accurate the data is? I think even 'accurate' data will show both Nvidia and Intel having more than 70% - Nvidia more than 80% almost for sure.
Isn't the bottom line clear anyways here - that people prefer almost-monopolies and 'big fancy' brands? Instead of them aiming for the GPU that's best for their situation and a healthy market?

P.s. for some an RTX4090 would be best, but for many others sometimes an Nvidia, sometimes an AMD GPU is best. Not this ~85% (Nvidia) vs ~15% (AMD) nonsense.

 

[LimyG]

But does it matter how accurate the data is? I think even 'accurate' data will show both Nvidia and Intel having more than 70% - Nvidia more than 80% almost for sure.
Isn't the bottom line clear anyways here - that people prefer almost-monopolies and 'big fancy' brands? Instead of them aiming for the GPU that's best for their situation and a healthy market?

P.s. for some an RTX4090 would be best, but for many others sometimes an Nvidia, sometimes an AMD GPU is best. Not this ~85% (Nvidia) vs ~15% (AMD) nonsense.

I think you may be taking this a it too seriously.
No one really knows, and I for one don' t even want to know.
It is just a light hearted thread with sarcastic jokes.

Finally, yes, the figures you quoted, "85% Nvidia and 15% AMD truely is nonsence.
It's so silly all one can do is have a laugh. I didn't notice the GPU share. Seriously that is the worst one. Surprised Valve even publish that.
Well, IMHO they know it's crud, but just post it because they know most peoples jaws will drop at first read incredulous. Then folks will just laugh at the daft figures. What else can one do.
BTW: please don't be offended, that's not something I want ot do at all.
I gave rep to your post as you will, and know that it's just light hearted. Cheers.

 

[EndRessentiment]

I think you may be taking this a it too seriously.
No one really knows, and I for one don' t even want to know.
It is just a light hearted thread with sarcastic jokes.

Finally, yes, the figures you quoted, "85% Nvidia and 15% AMD truely is nonsence.
It's so silly all one can do is have a laugh. I didn't notice the GPU share. Seriously that is the worst one. Surprised Valve even publish that.
Well, IMHO they know it's crud, but just post it because they know most peoples jaws will drop at first read incredulous. Then folks will just laugh at the daft figures. What else can one do.
BTW: please don't be offended, that's not something I want ot do at all.
I gave rep to your post as you will, and know that it's just light hearted. Cheers.

The 85% vs 15% is what I saw from a source (not Valve) I found when I did a quick google on the market shares.

Thanks for pointing out that I was taking things too seriously , it's only logical given the things I had to deal with (that sounds very serious too, I know! Bear with me, please )
But it's certainly not necessary to take things so seriously. It's good to have a laugh, shrug things off.

 

[LimyG]

The 85% vs 15% is what I saw from a source (not Valve) I found when I did a quick google on the market shares.

Thanks for pointing out that I was taking things too seriously , it's only logical given the things I had to deal with (that sounds very serious too, I know! Bear with me, please )
But it's certainly not necessary to take things so seriously. It's good to have a laugh, shrug things off.

Good stuff mate. Thanks for the reply.

 

[HardReset]

The 85% vs 15% is what I saw from a source (not Valve) I found when I did a quick google on the market shares.

You did it wrong then, there you go:

Btw that middle one is Not Nvidia.

 

[LimyG]

You did it wrong then, there you go:

Btw that middle one is Not Nvidia.

Wow, so according to that graph the new (and getting better rapidly with improved drivers) has Intel as the top GPU seller.
While I do believe and hope that Intel will keep up the pace, soon we will have a third choice at the top end. Impressive.
But those sales figures are even worse than Steam's lol.

Intel has had over 60% of GPU market share for the past year!?!
I suppose you are joking, but really those figures are utter rubbish.

Nice one. Pull the other. Ha Ha.

 

[HardReset]

Wow, so according to that graph the new (and getting better rapidly with improved drivers) has Intel as the top GPU seller.
While I do believe and hope that Intel will keep up the pace, soon we will have a third choice at the top end. Impressive.
But those sales figures are even worse than Steam's lol.

Intel has had over 60% of GPU market share for the past year!?!
I suppose you are joking, but really those figures are utter rubbish.

Nice one. Pull the other. Ha Ha.

Those are indeed accurate. CPUs and motherboards containing integrated graphics are counted there.

While those usually are rubbish, many of them are better than Nvidia's low end trash like GT710.

 

[LimyG]

Those are indeed accurate. CPUs and motherboards containing integrated graphics are counted there.

While those usually are rubbish, many of them are better than Nvidia's low end trash like GT710.

Including integrated graphics. Well that changes everything.
I was thinking discrete GPU only. The numbers make senses now.

I hever use integrated graphics of course as a gamer, but as you alluded the more recent chips actually have decent performance. (I only know that from my habit of reading lots of hardware reviews.
Any how, thanks for casting light on that. At first I thought WTF!!
Now I get it. Cheers.

P.S. One thing I think you will agree with is that it's just daft to buy modern i5, i7, i9 and then get a super cheap useless GPU. I'm still on a a i9-9900Ks but it's great and OCs well.

If I wasn't a gamer I would never buy a GPU with a chip like that.

 

[waclark]

With all the inconsistencies pointed out above, it turns out that this data set isn't all that reliable...

Inconsistencies does not mean the data isn't reliable. It just means that we don't know why certain stats went up or down unexpectedly.

 

[bexwhitt]

Steam survey is *highly* suspicious. This month the percentage of Chinese speaking users went up by 13.71 percentage points. All of a sudden they added tens of millions of new users and only in China!? That happened a few months ago, with many inexplicable large jumps in various categories, only to be completely reverted a month later.

probably party animals that's massive in china and on steam.

 

[yesnt]

Maybe the Great Chinese Firewall accidentally let all this info through this month lol

Maybe there are people outside of China who speak Chinese (whichever version).

 

[LimyG]

Maybe there are people outside of China who speak Chinese (whichever version).

True, hadn't considered that. Very approx Chinese make up almost 1/4 of the worlds population. (It was 1/3 when I was a kid, but that's back in the olden days of Pong).

Most large cities, and especially Vancouver or Canada as a whole have very large populations of Chinese. So yeah, could be part of it. On the other hand it's been like that for a decade or so.

@ Bexwitt, what is this game you speak of, "party animals?"
Sound pretty cute unless it's actually an animal FPS with gore all over the place. That would be fun.

 

[Mokona]

Steam survey is *highly* suspicious. This month the percentage of Chinese speaking users went up by 13.71 percentage points. All of a sudden they added tens of millions of new users and only in China!? That happened a few months ago, with many inexplicable large jumps in various categories, only to be completely reverted a month later.

It's not suspicious at all, it's a voluntary survey that any user can take, and neither claims nor tries to be representative of Steam user population in general. It's kind of just for the hell of it.

If you want an accurate picture of what people are using, you either need to capture the values for a carefully selected representative sample group that can be generalized to the entire user base with a relatively low error rate, every month, or Steam simply needs to put having your hardware, OS and language settings profiled into the ToS, and get the entire data set. Either option of probably more trouble than it is worth for them.

So we get a non-representative opt-in survey instead, which really only tells you what the people who decided to take the survey reported. That data is not generalizable to the full userbase with even a remote degree of accuracy, and can't, and shouldn't, be used to infer anything about that population.

 

[Maxiking]

It's not suspicious at all, it's a voluntary survey that any user can take, and neither claims nor tries to be representative of Steam user population in general. It's kind of just for the hell of it.

If you want an accurate picture of what people are using, you either need to capture the values for a carefully selected representative sample group that can be generalized to the entire user base with a relatively low error rate, every month, or Steam simply needs to put having your hardware, OS and language settings profiled into the ToS, and get the entire data set. Either option of probably more trouble than it is worth for them.

So we get a non-representative opt-in survey instead, which really only tells you what the people who decided to take the survey reported. That data is not generalizable to the full userbase with even a remote degree of accuracy, and can't, and shouldn't, be used to infer anything about that population.

Bruh, that's not how it works and you have no clue.

1. Every survey includes people who DID DECIDE TO OPT IN and is made voluntarily.

2. Every survey has some degree of an error, higher or lower.

3. You do not need a carefully selected sample either in this case, you are not polling 1100 people in an attempt to figure out the political preferences of the entire nation consisting of millions.

4. All you need in this case, is a RANDOM selection of the people who will opt-in and a big enough sample, let's say 50k people+ and you do not even need the unique entry either. And the potential error is a very simple formula n/N.

And yes, the steam survey is pretty representative, it is updated every few months, it correlates with the trends on the market, every time Nvidia releases its GPUs, the survey is quick to react and we see the wide adoption, same with Amd or Intel cpus when they are hot.

 

[bondflavor]

Bruh, that's not how it works and you have no clue.

1. Every survey includes people who DID DECIDE TO OPT IN and is made voluntarily.

2. Every survey has some degree of an error, higher or lower.

3. You do not need a carefully selected sample either in this case, you are not polling 1100 people in an attempt to figure out the political preferences of the entire nation consisting of millions.

4. All you need in this case, is a RANDOM selection of the people who will opt-in and a big enough sample, let's say 50k people+ and you do not even need the unique entry either. And the potential error is a very simple formula n/N.

And yes, the steam survey is pretty representative, it is updated every few months, it correlates with the trends on the market, every time Nvidia releases its GPUs, the survey is quick to react and we see the wide adoption, same with Amd or Intel cpus when they are hot.

Voluntary Response Bias in Sampling – InMoment

Voluntary response bias occurs when you allow your sample to self-select, resulting in skewed data that isn’t representative of your whole population.

inmoment.com

What is a Voluntary Response Sample? (Definition & Example)

This tutorial provides an explanation of voluntary response samples, including a definition and several examples.

www.statology.org

Voluntary sampling is statistically flawed if meant to extrapolate to an entire population.

 

[Top S]

A bunch of Chinese net cafe machines with Intel and RTX 3060 resurfaced due to some polling mechanism change?