Originally posted by RussianNeuroMancer
View Post
Announcement
Collapse
No announcement yet.
Steam's September Hardware Survey For Linux
Collapse
X
-
I got the steam user survey for the first time this month, I have been using steam since the earliest linux betas. I suspect this is a random sampling thing. I don't think the sample numbers are far wrong. Many linux users either dual boot or use wine because the games they want to play are not available. I have avoided doing that, but I don't think it is all that common.
Comment
-
Originally posted by Bucic View PostThis alone should end this bullshit. Doesn't Michael know the polls are bull?
Canceling the poll and changing the OS is probably detected, to keep people from willfully manipulating the poll by changing to their OS of choice before taking the survey. I would say what is being reported as a bug is actually a feature to prevent people from "cheating".
Comment
-
Originally posted by dh04000 View PostI run windows and linux and I got polled on Linux this time. Its not baised, you get polled on whatever OS your on at the time you get selected to poll. I spend >90% of my time on linux, so it makes sense that it was very likely that I was on linux at the time of polling. Statistics.
Why do people call things biased if they don't agree with them? Its like dealing with the climate change deniers.
They should remove all ambiguity and just collect the OS and hardware string on login for every user over the last 30 days. Then just go with what was used most often over that month.
Comment
-
Originally posted by xeekei View PostWhy does it "randomly" select people if it's supposed to survey how many people are using what? It should ask everyone.
Sample Size: How Many Survey Participants Do I Need?
What is a Representative Sample Size for a Survey?
Comment
-
Originally posted by Wildfire View PostBecause randomly selecting a large enough sample produces (almost) the same result as asking everyone (e.g ?1%). However, it is a lot faster and produces a lot less data. This is called a "representative survey". For some it might make a difference if the percentage of Linux gamers is 0.9%, 1% or 1.1% (or even 2%) but ultimately it doesn't matter. Who cares if Windows, in total, owns 95% or 94.9% of the market. It's undeniably the most used platform so publishers will continue to target that platform. Don't fool yourself thinking that Linux would suddenly have a much larger share (e.g. 20%) if only the survey where "more fair".
Sample Size: How Many Survey Participants Do I Need?
What is a Representative Sample Size for a Survey?
Also, with 50 million users its quite possible for a random sample of even 5 million to return 100% of users using Windows.
Thus it's best to remove all ambiguity and get everyone's data that has logged in via their client over the last month.
Comment
-
Originally posted by Kivada View Post[..]
Also, with 50 million users its quite possible for a random sample of even 5 million to return 100% of users using Windows.
[..]
If you have 1% Linux users, the probability to draw a Windows user is 0.99 = 99%.
If you draw two users, the probability for both to be Windows user is 0.99^2 = 98,01%.
If you draw 1000 users, the probability for all of them to be Windows users is 0.99^1000 = ~0,0043%.
For 5 million, it's ~0% and not "quite possible"...
Comment
-
Originally posted by uid313 View PostAny 64-bit Steam client yet?
Any 64-bit Linux games on Steam yet?
Any open source games on Steam yet?
heres a 64bit game
Comment
Comment