I tend to use relays, because if there are more capture points than one, you can just trigger that relay from all of those instead of implementing all stuff into each one. Also you can have different relays for different teams if you want different things to happen upon each team capture.
I figured it out. Whenever a point is capped, both Logic A and B are triggered. Logic B is disabled on start. Both logic functions will disable themselves and enable the other when triggered, as well as enabling the tracktrain to go forwards and backwards.