Worth mentioning here are differential flexure, wandering mirrors, weak focusers and other problems: for any guidescope to work, it is of paramount importance that there's no relative motion whatsoever between the guiding optical system and the guided one. If you cannot tame your optical setup, then the only option is an off-axis guider, but it's a hard way...
This problem becomes more noticeable the longer the imaging telescope's focal length,but in no case a longer guide scope is the solution.
Why is it best to guide with a short scope?
Autoguiding is one of the most challenging tasks to master in astrophotography; nowadays, with digital cameras (CCD or CMOS), manual guiding makes no sense, as it is almost impossible to keep the required accuracy, even more with the ever-increasing exposure times everyone is using.
There are quite a lot of misconceptions regarding the ideal scope (and camera) for guiding; enlarging the focal length, sometimes even adding barlow lenses, is probably one of the most popular.