Speaking of Network stuff... Can anyone explain why the Japanese use 85 snaps? Would setting packets to 85, snaps to 85, fps to 85, provide a better response from the server? Also, how could this effect the 'magic number' settings? (read the Upset Chaps Q3 guide to FPS and netcode if you don't know)

setting such a high snaps value is nonsense. they probably dont even know why they're doing it (mostly 1 starts and then others blindly copy it). the snaps value is also limited serverside by the command sv_fps which is mostly 20 (shouldnt be higher than 30) in ET. setting snaps any higher doesnt have any effect. and i doubt that they have sv_fps higher than 30. this setting shouldnt really be modified, as high values will increase the bandwidth used by the server and also fukc up engine related physics,...
using a lower snaps setting than the sv_fps on the server on the other hand will cause problems for the client as he does not request/send the amount of updates necessary for smooth prediction/gameplay (you'll notice your lagometer is getting jerky). i recommend to never set snaps below 20, i use 40 atm which is not really necessary on our servers(using same cfgs for et/tce and there are some et servers with sv_fps 40) but a higher value wont do any harm.

conclusion: looks like (based on baal's observeation - i never checked tbh) some japs are trying to be extra smart or just cant get enough of placebos (cg_skill 1, makes u pwn!)
remember: /snaps 30+ is ok, <20 is *I need a large hot beef injection ♡*

Hmm i though it was 45
My fps is jumpy so i might be that welcome on servers since my fps jumps between 2 and 90

- cl_maxpackets should be equal or half com_maxfps.
- snaps should be equal sv_fps or half sv_fps for some smooth packets transfer.
- if snaps is half sv_fps then cl_packetdup should be 1.

The convention in quake engine was to get sv_fps to 20 or 30, thats why snake is right when he says having snaps to 40 will always make you fit the situation.
Dont know which effect would give sv_fps 85, i did some secret tests on FORCE Main server some time ago, and noone found any change (better or worse).

After some tests @ timenudge i found it really helped a lot when being on servers with high ping (like jap servers) ; thats why i say it could be useful to re consider the restriction about this cvar.
Pbbans recomendation (in MCI) : pb_sv_cvar cl_timenudge in -20 0

We are offtopic btw