Steam now allows you to install any Windows game on linux in the same way you'd install any other game. That's a big improvement.
Anti-cheat is an issue, but otherwise it isn't a million miles away.
The value 88(%) is also misleading. It represents games running at least at bronze level (the lowest quality of a playable game). The percentage of perfectly running games (via Proton) is 15%.
I support Proton as a shortcut to release Linux-compatible games, but there's no comparison, in the big picture.
I'm less about trying to imply that Linux is "crushing" Windows and more about it being seen as a viable option (i.e. " not a million miles away").
I think I'm against out of hand dismissal, which your post and the post I responded to are without qualification or balance.
I didn't intend "shortcut" as derogative (bad word choice :)); rather, given the limited resources of game development, it may make Linux ports possible (as opposed to native ports, which take significantly more resources). John Carmack actually supports this approach.
My criticism is that while there is the impression/hope that at some point the "emulation" will be completed, based on my experience (with Wine and ReactOS), it never will - it's a virtually infinite amount of new hacks/tweaks/small additions.
My conclusion is it a type of work that will require continuous employment of (development resources), and will be more of a workaround, than a solution. (regarding the "crush", this was actually in reference to an ancestor post)
Including native ports in the discussion, or excluding them, is a tricky discussion, which I consider separate - my points were about Wine/Proton specifically.
I actually do play games, but even if my O/S is Linux (and essentially all the games are on Steam), I still run games in a VM.
Markets don't optimize for efficiency in computing.
I do think the above poster hasn't made it clear the experience they're speaking to. A lot of "anti" sentiment I see is based on things that aren't really true anymore. If all were informed and were giving their considered opinion that despite recent advances they still don't foresee a large gain in marketshare then that's fair enough. I think my main issue is with uniformed generalisations being stated as if fact backed up by recent and relevant experience.
Having said that I do agree with your post, so please note that my criticisms don't apply to you!
The "news" is that in the last 5 years it has gone from "absolutely not viable for gaming" to "good enough for a lot of people".
OS/2 was also a great platform for games by that measure, thanks Win-OS/2.
Same applies to GNU/Linux, hence Windows continues to thrive, while the 1% of the desktop folks get to taste Windows games with an ABI emulation layer.
Linux is mostly not being used because people don't have all the software they need and translation layers solve that.
We're entering a different space in desktop OS usage. Windows is changing the desktop metaphor to be more similar to macOS and mobile computing is being used by an increasing number of people as a daily driver.
In short, I don't think ABI emulation layers are the reason Linux is only at 1%. Indeed it was at 1% before Proton was released.
I think "it does not play most recent games" is untrue without qualification.
It is however one of todays most well known games.
I'm not diminishing the value of games like that running well under Wine, I think it's great. But it's not the same as "recent" games running well.
As one of the GPs points out anti-cheat is unfortunately another big hurdle. CS:GO works great because VAC runs in user-mode and Valve go out of their way to specifically detect Wine and adjust behavior as appropriate (I reverse engineered it extensively some years ago). I think EAC has a native Linux client, but many of the big names are still unplayable.
If you try to play Valorant, Apex, Siege, Warzone, PUBG, etc. my understanding is they'll barely even launch, let alone allow you to join a multiplayer match.