1. Load some large terrain ( 50 km x 50 km )
2. Put armed soldier in NE corner
3. Observe his rifle in 3rd person view ( when rifle is put down using Ctrl-Ctrl and soldier is playing idle animations )

Or :

1. Load a mission attached in a file Jittering.VR.zip

( soldier is placed in position (47k,5,47k) ).

2. Observe his rifle in 3rd person view ( when rifle is put down using Ctrl-Ctrl and soldier is playing idle animations )

Or :
You can also look at this video :
http://www.youtube.com/watch?v=wug0pnApRys#t=201
It is particularly visible from 4:00 to 4:19.

In short : object positions are (usually) represented using
floating point data types ( google for IEEE-754 on wikipedia ).
The thing with floating point data type is that when the
numbers are big, they lose precision.

In computer graphics this starts to occur when object position coded in single precision floating point is more than 30 000 units ( usually meters ) from (0,0,0) point.
And it starts to look like the position of the object is "quantified".
Look at the before mentioned video at 4:05 - the rifle looks like
it moves for a certain number of centimeters every frame - its
movement is quantified.