The relay component switches to the intermediate state immediately giving incorrect results for some circuits.

[Glidos]

I was testing a circuit built from many relays and diodes. The simulator uncovered an unexpected failure case where the back emf from the coil of one relay activated another. I was (and still am) impressed at the simultor finding that behaviour, but have realised that the behaviour would very likely not occur with real relays: the transients from the switch off are only momentary, too short a period to cause any significant movement of the switch. It would be useful to be able to specify delays for both parts of the switching action.