This thread is tenfold more interesting than the purely technical opinion on Techspot it linked. It’s like hitting a string of one instrument, and setting the entire orchestra pit in resonance.

There are 4900 issues open at the proton's github page.
And how many issues are there open in some alternative cases? Oh, I forgot: you can’t see them. And if you can’t see something, the count must be 0! /s
To give missing context for the number, Steam now offers about 100,000 games. Feel free to compare 4,900 either to that, or to even a modest estimate of total number of issues in these 100,000 games. In either case the ratio is lower than the value Daniel Sims gave us (himself
taking that from Alfonso Maruccia).
There is much more people than you may expect, who would be happy to adhere to the openness ethos and to make their organizations coöperative and honest. A simple observation and a calculation even a fool can make convinces them it’s a bad idea. They’d expose themselves to attacks like that. So in fear of getting hurt, they choose not to.
But for those who wish to finally abandon that Trojan Horse called windows, and held back due to games, give it a go. You will be both liberated and pleased.
There is no reason anymore to be owned by that other garbage ad-and-spyware-infested, subscription-pushing OS for games or anything else.
Move to operating systems based on Linux and… turn them into a similar pile of poop? To centralize the market even more, now Linux based systems being the next platform one is forced to use?
What Windows ecosystem is, is not because some CEO’s decision. It’s a global, decades-long process encompassing entire societies. Changing a kernel is not going to stop that. Use Linux, XNU, Mach, a BSD, and the same process is going the happen, because the people are the same.
Ask ten linux users which distro is best, you'll get ten different answers and therein lies the rub.
It’s the same as with any other product or service, no?
I'm not much of a gamer, but wasn't there cases quite some time ago of certain Steam games running even faster on Linux than they did on Windows?
As it should be, shouldn’t it?
It’s a probabilistic distribution, so some values will be above, some will be below. The important thing is how the entire distribution is shaped, not if there exist single instances that are better (or worse).