Comment Story that didn't print important thing last (Score 1) 49
Story notes that number changes massively every month and it went 3,5 > 2,5 > 5,3 across January, February and March this year.
This obviously isn't changes in user base, but changes in tracking combined with very low sample size, meaning it's wildly inaccurate.
Props to writer in that he didn't engage in mainstream media clickbait method of "we clickbait headline and then we spin a narrative for the entire story. And then we destroy our narrative by telling what actually happened in last two paragraphs". This is explained in three opening paragraph, and literally first sentence is "if figures are accurate" proceeding to explain why they almost certainly aren't:
Quoting the story's opening below:
"If Valve's latest Steam Survey monthly figures are accurate, Steam on Linux enjoyed a very wild month of March. Steam on Linux is now above the 5% threshold and more than twice the size of the Steam on macOS marketshare.
Steam on Linux ended 2025 at around a 3.5% marketshare, dipped a bit in January, and fell to 2.23% in February. That's still much better than several years ago in the pre-Steam-Deck days when Steam on Linux was at around 1%. In absolute terms with the continued growth of the Steam user base, 2~3% was rather healthy considering all of its bumps over the past decade.
But Valve just published the Steam Survey results for March 2026 and they have never been so incredible for Linux... 5.33%! Steam on Linux was never above 5% and easily an all-time high for the Linux gaming marketshare, especially in absolute numbers. It was a massive 3.1% spike in March"
I.e. this growth is more than total supposed user base last month. This is obviously a massive statistical inaccuracy just hitting the limits of the error margin, as these numbers were tracked for many years, so we know what approximate number is. Error rate for these sorts of studies is usually 2-5% depending on sample size and methodology, so we're seeing the error rate manifest itself on massive differential reported.
It would really help if people took any decent class on statistics.
