We all know that the Steam hardware survey literally means nothing at this point, right? The only company that might have more or less accurate statistics about actual PC hardware trends is Microsoft.

Steam survey "only" matters regarding PC gaming because everyone that games on a PC probably has a Steam account. For a complete picture about all PC hardware available right now, only MS can provide it, like you said.

The survey is optional and it's actually not requesting info from users every month. This time (as others mentioned), it also seemed that DX8 GPUs were on the rise :infinity:

They can sit there with reasonable accuracy and say "The vast majority of steam users have less than 4 cores, thus we shouldn't make a game that absolutely requires 4 cores for an enjoyable experience" or "The vast majority of steam gamers are using 1080p and below" or "The vast majority of steam gamers have a 1060/480 or weaker".

I think that Steam survey provides a good picture of the actual hardware gamers are using right now but that´s my opinion, nothing else. If you want to dismiss it that´s up to you, no problem.

All true but do you know any other company that provides all this info like Valve does? Or any other company, with the exception of MS, that can provide info like this correctly? I don´t. Also do you think other stats provided by Steam, like core counts, most used monitor resolutions and Nvidia/AMD market share are also wrong? I think that Steam survey provides a good picture of the actual hardware gamers are using right now but that´s my opinion, nothing else. If you want to dismiss it that´s up to you, no problem.

No they can't. If it only compares to the previous months data, that is even worse. As i said, this information does not tell anything useful to anyone. It has no actual information in it other then "Well this is what was found this month, in an extremely small selection of devices, that does not represent the whole, or represent any market trends due to not testing the same previous steam accounts from previous surveys and etc." Again, as a developer, this is completely useless information. How can it, when it does not survey anything useful? How can you know what people are using "right now" if you decide to only survey random people that do not present the whole? That does not work. You'll get what SOME people are using right now, but you'll get no accurate information on what MOST people are using right now or market trends (which are the most important thing when making software/games is to see WHERE people are going)

No they can't. If it only compares to the previous months data, that is even worse. As i said, this information does not tell anything useful to anyone. It has no actual information in it other then "Well this is what was found this month, in an extremely small selection of devices, that does not represent the whole, or represent any market trends due to not testing the same previous steam accounts from previous surveys and etc." Again, as a developer, this is completely useless information. How can it, when it does not survey anything useful? How can you know what people are using "right now" if you decide to only survey random people that do not present the whole? That does not work. You'll get what SOME people are using right now, but you'll get no accurate information on what MOST people are using right now or market trends (which are the most important thing when making software/games is to see WHERE people are going)

The survey is optional and it's actually not requesting info from users every month. This time (as others mentioned), it also seemed that DX8 GPUs were on the rise :infinity:

Almost all surveys are optional and don't request from the same people every month. That's how good surveys work, random sampling. As for DX8, the amount of DX8 GPU's that were polled probably increased. The sampling is random and has some degree of deviation which would be higher with DX8 GPU's because there are so few of them.

People that have alleged "problems" have a problem it's not Windows 10.

If you run only Windows 10, install only ms store apps, do not change a thing in system and not use any specialized 3rd party programs then Windows 10 runs very well... But if you want start using Windows 10 like you used Windows 7m run your favorite antivir, firewall, change security options, run specialized programs then you starting have problems, problems what you not have in previous system. So some peoples do not like Windows 10 and going back, because they want do more than just run os itself...

No issues with all the above. I uninstalled the Store and Xbox App, installed NOD32 like i have for the past 10 years, installed the same software as on Windows 7. Changed some settings in Group Policy, disabled some services like i did in Win 7... It all works fine. You're just looking for problems that arn't there.

Feel free to keep invalidating peoples real world experiences because you haven't had them.

Good surveys don't really have "random" sampling. They are properly distributed actually, and they also come with error tolerances in percentage form. None of this seems to apply for the Steam numbers we get.

And sadly huge increase in 2-core CPUs at expenditure of 4-core CPUs... Yeah, intel mobile i3=i5=i7. All dual cores with HT.