My experience is that the performance of videos depends a lot on how much of the screen the video takes up. More so than on the quality setting (480p, 720p, 1080p). Usually when I want to watch a video, I limit it to about 1/4 of the area of my 1920x1080 screen. But YouTube is pretty watchable. I don’t think even when you select something like 720p@60fps or whatever that it ever actually gets up to 60fps, and sometimes the parts that don’t compress as well drop a few frames (strobing in particular does that), but it’s really not bad if you’re ok with not doing full-screen.

Also, it depends where the video is hosted. If the JS is really light on the page, it’s way more performant. If the JS is really heavy, I sometimes have to wait tens of seconds for it to render the first frame.

But also, this Raspberry Pi is mostly my Lemmy box. Not my main box.

Exactly, the stack necessary to play videos in your browser is no joke and I’ve always used it as a litmus test.

It’s not fool proof – as you pointed out – it’s probably the closest thing to a litmus test there is