Just wanted to leave this here for awareness.
First this take about using ProtonDB - https://boilingsteam.com/ubuntu-and-manjaro-fall-further-into-oblivion-february-2024/
Secondly the Steam survey - https://store.steampowered.com/hwsurvey/Steam-Hardware-Software-Survey-Welcome-to-Steam?platform=linux
Compared to 2019 for example, things have changed A LOT.
Iām fairly certain a lot of people, especially those on the Desktop and Gaming teams at Canonical are aware, but I think itās a bit disingenuous to just drop information without coming up with at least any speculative reasoning or possible solutions. Thatās part of what community contribution is all about. 
With that in mind, why do you think this is? Do you have any solutions to this problem?
The problem with a 100% stacked bar chart, as used in that article, is that it gives no information about absolutes - is the total number of users represented by each % point shrinking, stable, or growing?
Ex. if Ubuntu was 30% of 1M four years ago, and is now 15% of 10Mā¦then the absolute number of Ubuntu Steam users has skyrocketed higher, but a 100% stacked bar will make it look like āfalling further into oblivionā (what a clickbait title).
I suspect that chart was used because the absolute data isnāt actually readily available, but IMO thatās the first question - are people actually switching away from Ubuntu Desktop? And if so, is it for a reason thatās relevant to and solvable for Ubuntu overall?
Iām far from being an expert in the field (my idea of āgamingā means, essentially, chess), but looking at the data between 2022 and this year, Iām getting the impression that the only growing options are specialized distros (SteamOS; Garuda).
Ubuntu is not a good fit for something like that - unless someone spins up a gaming flavour. 
I donāt have an idea about it nor a solution. Just letting the community be aware of it.
The first chart seems to be missing SteamOS / Steam Deck. According to Valveās stats that should occupy around 42.12%.
I also agree absolute stats would be helpful to show if Ubuntuās decline is actually growth.
This is a VERY important point right there. I am more concern, not so much the switching, but the idea of thinking Ubuntu is not the first distro someone would think about when also thinking about gaming. Like a previous post I made where, with just 1 or 2 tweaks, Ubuntu ran laps compared to others. But it is indeed worrisome those articles (like this one), when the point about relative and absolutes you mention I do not think was analyzed for that graph.
Yes, that absolute part I did not even think about it. The graph might be misleading and scary at the same time to anyone trying to migrate to Linux and thinking something is seriously wrong with Ubuntu.
My perspective for whatever itās worth: I wouldnāt stress about it. IMO a person has to be way deep into the āLinux enthusiast userā space to even make it to a site like the one to which you linked - the first ten hits when Googling / Startpaging the phrase āSteam on Linuxā are mostly distribution-agnostic, informative or instructional, but the two distributions mentioned directly in those link titles are Ubuntu and SteamOS.
I would be much more concerned about things that can be done practically so that when folks do arrive on Ubuntu, and are interesting in gaming on Linux, they have a good experience (e.g. testing/reporting Steam Snap or Ubuntu packaging bugs that impact gaming).
To be honest, I think the whole āproblemā here is a tempest in a teapot and is no problem at all. If Ubuntu is declining but Linux overall is increasing, I donāt see a problem since those games will just work on Ubuntu anyhow.
Therefore, in my opinion, considering thereās no problem, no solution is needed either. Linux is up because SteamOS is up. If Linux and SteamOS are up, weāre winning!
Itās about seeing the positive rather than the negaitve. No reason to stir controversy when there is none. 
Just to chime in with my first post on the Ubuntu Discourse.
This is largely if not 97% because of the Steam Deck, since the OS is based on Arch.
Also if things are working as expected on Arch, having them work on Ubuntu is a sure thing. No matter what some people say, for better or worse, Ubuntu is still the most used distro/OS in the Linux world.
Author of the chart here. ProtonDB stats do not contain much data related to the Steam Deck, because most users who report their data on ProtonDB actually game on desktop or report compatibility using their desktop OS. So Steam Deck is a non-factor for this data set.
You can see the number of users who report their configuration on every month at the very top of the chart. Itās small, but should be readable. Since these are voluntary reports, itās a sample of the whole user base, but itās fairly stable in terms of sample size (it has not grown that much over the past 5 years).
Nope, as mentioned in my other post. The Steam deck is not included at all in this data set, due to the nature of ProtonDB reporting.
In my experience, I expect that Ubuntu will have a relatively low percentage of people who are going to submit bug reports or game reports on ProtonDB. That doesnāt mean that there arenāt a LOT of people using Ubuntu.
I have moaned at Boiling Steam in the past about their poor interpretation and sensationalisation of these stats. Itās as bad as Distrowatch. But they donāt listen to facts.
Ubuntu (and Manjaro) have a significant number of users. Many of those users play video games. A large chunk of those play games via Steam. A vanishingly small proportion of those users contribute in any way to anything, let alone specifically ProtonDB.
Itās poor journalism from a partisan author who has a history of writing clickbait.
Thank you, we appreciate your input for the data.
Yes distrowatch is a very old example on how to do bad fact checking.
The Steam HW survey is also a difficult one to refer to because Arch/Manjaro include all their users in one entry, whereas Ubuntu is split not only across releases but also major point releases. This means that the Ubuntu 22.04 LTS number that is visible does not represent the gamers on Ubuntu 23.04 or 23.10, Ubuntu 20.04 or the Steam Snap (which by itself is around 90k monthly users). All of these Ubuntu gamers are categorised as āotherā which provides a misleading impression of how Ubuntu compares.
I did not know about the āOtherā part. I think steam needs to update the way they gather the information for this kind of things. If this is happening to Linux distros, there could also be some other quirks going on with the rest of the stats.