AMD says Steam Hardware Survey miscounts processor data

[HEXUS]

Steam's algorithm is confused by multiple user logins in Chinese internet cafes it seems.

Read more.

[Corky34]

So the purpose of their data is to show general trends to game developers but they're not motivated to get things right, that about sums up Valve these day, we just can't be bothered.

[GinoLatino]

They should count/identify the machines, not users...

[LSG501]

As much as I like AMD and am sure the figures aren't accurate in relation to market share isn't the chart supposed to show what is being used to play stuff from steam.

So seeing as it's basically using 'user logins' to count the cpu being used then it's kind of accurate even if it's not accurate in terms of physical hardware being bought.

[Ttaskmaster]

They should count/identify the machines, not users...

Machine ID is kinda the same idea as a vehicle's VIN, and so accessing/using that data might be considered personal info and/or contravention of GDPR or some such, no?

Basically, what right would Steam have to know your machine's unique ID?

[.havoc]

Could just use a similar implementation as storing a cookie or just setting a flag on the system as being checked in this quarter (or whatever the frequency is). Wouldn't need any further information than access to their settings...no?

[Hoonigan]

Machine ID is kinda the same idea as a vehicle's VIN, and so accessing/using that data might be considered personal info and/or contravention of GDPR or some such, no?

Basically, what right would Steam have to know your machine's unique ID?

Do computers even have such a thing?

[Corky34]

Basically, what right would Steam have to know your machine's unique ID?

Probably the same right Microsoft have of doing that, maybe not directly comparable but as .havoc points out you wouldn't really need a machine ID, just some sort of flag to mark if it's been checked.

[Ttaskmaster]

Do computers even have such a thing?

I was just about to mention Win10 and how it's tied to your mobo ID....

Probably the same right Microsoft have of doing that, maybe not directly comparable but as .havoc points out you wouldn't really need a machine ID, just some sort of flag to mark if it's been checked.

MS have probably got it legally justified under licencing laws, though.
What Steam supplies is tied to a user's account, not one specific machine, which is necessarily thus so you can sign in wherever you like. Also not sure about cookies or any flag thing like that, as the user would have to agree to accept it, which then makes it a more subjective measure, right?

[EvilCycle]

Hang on a second, aren't these results just from the steam hardware surveys where people allow steam to check their hardware? No algorithms or login data as far as I knew? And even so, these results are showing what is used on steam, not what is being bought, so these results are correct in any event right?

[Ttaskmaster]

No algorithms or login data as far as I knew?

How's the app going to check if you're not logged into it?

[ET3D]

As I see it, the Steam survey is pretty useless for game developers. It would have been more meaningful if it had some breakdown by region.

[DanceswithUnix]

Do computers even have such a thing?

Kind of...


Motherboards have an Ethernet MAC address. If correctly implemented that is a 6 byte unique number, but in reality there can be duplicates and you can re-program it with whatever you want.

Hard drives/SSDs have a unique serial number which you can view using SMART tools.

Intel did put a unique serial number in the Pentium III, but a privacy outcry meant they dropped the idea and no other cpu companies implemented it.

The BIOS in the motherboard has a serial number storage field, but that is set by the OEM. As I built this machine and don't bother setting it up, I still have the Asus default value of something like "Serial Number" stored in there.

Microsoft mash these and things like what sort of video card you have to create a system "fingerprint".


OFC for what Steam actually want from their stats what they have now is kind of correct as that is what developers need to target. It just doesn't track sales figures, and it mushes casual gaming laptops in with the high end gaming desktops.


Edit: Amazon give a snapshot of their sales ranking, it seldom ties in well with Steam changes.

https://www.amazon.co.uk/Best-Seller...rs_2_428655031
https://www.amazon.co.uk/Best-Seller...ers/430512031/
https://www.amazon.co.uk/Best-Seller...rs_3_430512031

[Xlucine]

Makes sense - it's a hardware survey for games developers, not hardware enthusiasts, so counting the number of steam users with X CPU is what developers need (i.e. Y % of people-who-buy-games use X CPU)

[HollyDOL]

tbh I can't see anything wrong with that... If 80% users played on Celeron, then well, 80% users played on Celeron.

The statistics is based on unique users, not unique cpus. So if you as a dev tune your game to play better on Celeron, you improve experience for 80% players...

[Stukes]

Agreed - it makes sense from the "for devs" point of view. Still, it would be nice to have a "per machine" option too. I don't imagine it would be particularly hard to implement as a feature on top of what's already there, and it would be interesting market-trend data. Not just CPU/GPU, but OS, Screen resolution, laptop usage and all that.