Ah, OK. I've not seen this done before and I am not sure why it is done. The orange lead on a 3-port has two functions. When there is call for heating but not for water, the valve moves to the far position and a contact completes a circuit from white to orange to feed the boiler / pump. The only function the capacitor might have in relation to this is as a suppressor to reduce arcing at the microswitch when switching the boiler/pump load. I would be surprised if the microswitch would fail rapidly without it, unless they have started using very weedy microswitches or the present generation of mid-position valves actually use a solid-state switching alternative to a microswitch (It's years since I looked inside one, perhaps someone could confirm?) Personally I would use a snubber with built-in resistor for that purpose and the only reason to add it outside the valve would be if the one built into the valve was inadequate for certain loads. Mention of the pump overrun would suggest that the switching device in the valve is being damaged by the pump motor being dumped back on the orange lead and creating a transient after running from the boiler PL via the overrun thermostat.
There is another sneaky function for the orange lead, which is to backfeed the demagnetisation resistor in the valve when there is call for hot water via the normally-closed contact in the cylinder stat. The small resulting AC current in the winding is supposed to avoid the situation where, if the valve has been sat at the mid-position for a long time with DC flowing through the motor via the stall diode, the stator becomes slightly magnetised preventing the spring spinning the motor back to the rest position once the call for heat via the white lead is removed. There is a minor side-effect if the valve is in the far position when this happens: The valve stays put because grey is still live and current flows via the demagnetisation resistor back out of the orange lead. I can see a possibility that if this is used to control boiler electronics via an input that is unduly sensitive, the 1mA or so from the orange lead might cause unwanted triggering. The capacitor would provide a load to absorb that, however it would not damage the valve so I don't see that this is the reason.
If anyone can confirm or deny whether modern 3-ports have a triac or something controlling the orange lead, that might shed some light on it.
