Let us not oversell Proton. While the compatibility story now is better than it ever has been, only about 26% of the top 100 games on Steam work without substantial tinkering or other problems. And most of that is going to be due to multiplayer games and developer/publisher anti-cheat obstinance.
Statistically speaking, the games the average person would most likely want to play likely still do not run, or do not run without problems or tinkering required.
Where are you getting that 26% from? PCGamesN reported, last year, that around 70% of the top 100 were supported. It's a different story for SteamDeck, which has further constraints due to its unconventional screen size; but even then, ProtonDB claims at least 48% of the top 100 are playable on the Deck[1]. In my experience, just about any game that didn't work well on my Deck does work well after I switch it to Proton Experimental.
From the ProtonDB dashboard. Top 100, ProtonDB click play ratings system. 26% of those games are tier 1 or tier 2 (no way to give a direct link here).
You could go by the old the medal system, which looks promising at first glance, but "gold" basically means "playable". Platinum is the real metric there, and can be understood to mean "plays as well as it does on its native platform". There are game issues in gold rated titles that would lead to mass complaints and refunds if they existed on the native platform. That is kind of my point. "Playable" is way too low of a bar.
Ah, I think you're being too critical then. ProtonDB's tier 3 rating is still playable, just with obvious issues; like poor performance, configuration required or occasional crashes. Including those bumps it to 59%, and I suspect many of them work fine with Proton Experimental.
But like you said, you think "Platinum" is the metric that matters. A good chunk of my library wouldn't be considered platinum on Windows 11. ;)
Point taken, but I am looking at this from the standpoint of someone who already has a considerable PC game library but is not a big techie (a pretty good stereotype of the average PC gamer). I could not in good faith recommend a Steam Deck to such a person without knowing more about the games they play.
If they spend most of their time in esports titles not made by Valve (anything by Riot, Fortnite, etc.) they would be sorely disappointed.
However, if they are big into emulation or indie games, the picture is a lot rosier. (And not just because that is an easier workload, but because it implies more knowledge on the part of the user)
I think that is the tipping point for Proton and the Steam Deck, when they can be recommended without qualification to the average gamer. We are not there yet and I do not want to see people burned by having their first exposure to Linux gaming being so close yet so far away for their use case.
FWIW, I know a few dozen people who have purchased a Steam Deck and not a single one of them seems to be disappointed; the consensus appears to be that it's causing people to experience their libraries with new eyes. That said, we're all game developers and I don't think very many play sports games; or would think to do so on a handheld, when the alternative is on a couch with friends.
Myself, I'm using mine to power through all the puzzle and adventure games that languished on my PC while my attention was elsewhere.
Statistically speaking, the games the average person would most likely want to play likely still do not run, or do not run without problems or tinkering required.