Yeah this works because of indexed access. The tradeoff is you get a limited number of spaces, and you require a much larger key binding surface area. Like how indexes take memory.
If your state is A1/B2 (two monitors A and B displaying spaces 1 and 2 respectively), and you're working in A1 and want to switch B2 to B3, how do you "focus" on display B before moving spaces to B3? If you just press mod-3 I assume you'd get A3/B2. Do you have to alt-tab until you focus a window that is currently in B and then mod-3?
In my xmonad setup only one window (and thus screen) has the focus, so I would need to hit the key that switches focus to my B monitor (for me that's Mod4+e, B monitor is Mod4+q) and then Mod4+3.
True about indexed access; some other window managers have natively built in "tagged workspaces" where windows have tags and the "chat workspace" is just "all windows that have the chat tag". You can (probably? I've never used them) bind numbers to tags permanently or temporarily but you wouldn't be limited to 0-9. I never bothered trying to copy that setup in xmonad because I never used more than 5 or 7 workspaces.
If your state is A1/B2 (two monitors A and B displaying spaces 1 and 2 respectively), and you're working in A1 and want to switch B2 to B3, how do you "focus" on display B before moving spaces to B3? If you just press mod-3 I assume you'd get A3/B2. Do you have to alt-tab until you focus a window that is currently in B and then mod-3?