Do you know what the sample size for it is? Probably in the millions i would expect. If thats the case then i dont think 0.3% is just noise, but it also doesnt really show anythimg of value.
Comment on Valve: Windows 11 market share on Steam drops to 41.61%
Technus@lemmy.zip 7 months ago
For those unfamiliar, each month, Valve randomly picks a subset of Steam users and offers them anonymous participation in the survey. … According to Valve, in March 2024, Windows 11 went down from 41.96% to 41.61%. On the other hand, Windows 10 gained what Windows 11 lost—its market share increased from 54.19% to 54.40%.
Eesh, I dunno. Those are quite small changes and could just be a result of statistical noise, especially given the random sampling.
Not exactly headline worthy, IMO, but it must be a slow news day.
unexposedhazard@discuss.tchncs.de 7 months ago
Technus@lemmy.zip 7 months ago
I can’t find anything on the sample size. It’s not mentioned on the official page, and all the other results are armchair statisticians arguing about it in various forums. I’m guessing they want to keep that data close to their chest.
However, Steam’s charts page shows a peak of 34M players online at once over the last few days. A few different sites suggest Daily Average Users are around 60M. Let’s call it an even 50M for the sake of argument.
What would a decent sample size be without generating overwhelming amounts of data? Say, 10%? So that’s surveying 5M users.
0.3% of 5M is just 15,000 users. What if the survey just happened to pick 15k fewer Windows 11 users this time? Is that so unbelievable?
unexposedhazard@discuss.tchncs.de 7 months ago
Yeah that is insanely unbelievable. Maybe a tenth of that is barely realistic but 15k of sample error would mean that valve is completely shit at selecting a good sample.
jacksilver@lemmy.world 7 months ago
Even if their sample size is large, you’re assuming that people who are willing to respond to the survey are evenly distributed across different OS. I wouldn’t be surprised if their is a slant to this data just because it’s based on an opt in survey.
GreyEyedGhost@lemmy.ca 7 months ago
Well, the random sampling makes it less likely it’s invalid. Random sampling is pretty much the standard for large groups. As for significance, that’s something I’d have to read more to know whether it’s high enough to truly matter. There are very well developed rules for this stuff in statistics.
Artyom@lemm.ee 7 months ago
It’s a small change, but the trend should be decisively in the opposite direction. New Windows OSs are almost always growing in the market relative to old Windows OSs. This means Windows 11 is not popular and Windows users are avoiding it.
rooster_butt@lemm.ee 7 months ago
I’m willing to bet it’s just the tpm requirement holding off growth. People aren’t buying hardware just to install win 11.
lauha@lemmy.one 7 months ago
This small of a change is probably within a margin of error.