By using the method described by dnarkosis, you can almost totally eliminate bad midi-timing. Usually when people complain about the timing of their sequencer/synth, it is caused by too many midi-events triggered on the same clock-tick. If you shift them a little, the timing improves drastically!

The Microscope function is essential to know what is going on in the sequencer, but you need a basic understanding of the sequencer to be able to use it.

StringsMan:
I often create rythms by recordang a loop of, say, 8 bars and the copying it. In this situation I use quantize while recording. Sometimes I also play the rythm in real-time and quantize it later. I also work this way on fill-ins.

If you record something in real-time it is best to quantize post-recording. You then first get a track that is "hand-played" as good as you can get it, but not quantized. Quantizing actually destroys the original timing of the track, so it is best to do this after the recording, so you can use UNDO if it turns out wrong. This way you can also experiment with different quantize-settings safely. By quantizing after recording, you can play around with it, without the risk of destroying the data. If it is a real demanding track, and I played it very good, I usually save it to disk before even starting the quantizing!

Stig