[e2e] 150ms - tolerable latency for quakeIII

[Joseph Ishac]

I've been more of an UT player myself cause I couldn't get Q3A to run on my 
system.  Anyway the concept should still be the same.

A couple of thoughts:

First, I'm a poor 56Ker as any type of broadband is unavailable in my 
area.  For me "ping" times below 200ms is unheard of (I never find a server 
with one lower), and I usually end up playing with 250+.  The reason I 
quote "ping" is that I think that the program reports figures that 
represent more than just delay and often include packet loss and other 
factors (play UT, hit F1 and watch your ping to see what I mean -- often 
the value is very volatile and hits astronomical values when experiencing 
loss).

Second, movement for any player is often very wild, consisting of jumps, 
strafing, super jumps, teleports, etc...  The whole point being to attempt 
to be harder to hit, and because of this, I would image it being very 
difficult to simulate/predict a players movement.  Also, a lot more goes on 
than just movement and firing weapons in FPS.  Triggered stage events (such 
as traps), player chatting, verbal commands, item occurrences/respawning, 
objective's location (such as a flag), etc...

Third, a player never shoots directly at a target, but rather where a 
target could be.  As your connection speed increases you can compromise 
this somewhat, but it never completely goes away.  I shoot into empty 
hallways or lob grenades into a room without knowing if anyone is in there 
on many occasions and have noted many other with similar tactics.

Finally, what really annoys me is player warping (discussed in the XvT 
article and present in UT as well).  To suddenly "warp" in front of an 
opposing players rocket or over an abyss is not considered fun to me, and 
is what would cause me to drop out of a game.  However, warping seems more 
associated with packet loss as opposed to delay.

Just a few of my thoughts :)

-Joseph Ishac
NASA Glenn Research Center
jishac at grc.nasa.gov