Ok, now stop doing whatever it is that they're doing to the cursor layer that makes it feel like garbage (wlroots is especially bad, KDE less so but not as good as either Xorg or Windows, GNOME too but has other cursor issues so...) and then I'll finally consider daily driving any of this stuff

will try again with a photodiode instead, since it's known to be a valid way to measure stuff like this, and it seems precise enough at that

if it isn't this, then I'll probably have to dig into libinput or something

You gotta do the measurements. It's probably not even that hard, all you need is a USB mouse emulator (any microcontroller with USB peripheral support can do this and there are tons of examples) and a photodiode.

will absolutely do this, the microcontroller and mouse emulation part is solved for me already so I just need to get an appropriate photodiode and... profit

The problem with this is more so that I would still have to record this "IRL", as recording it in software is just... meh (I mean, I could try), and I just don't have the equipment for that (I demonstrated as much there), so I will probably end up doing the Pico + light sensor thing as described there. Should be more reliable anyway (something very similar has been done already, and successfully), so... yeah, that.