Arrgh... I can't stand all this well intentioned but inaccurate information. Here's one more opinion. Add it to the above posts and choose the one you like.

Twisting the wires is effective only in the following two cases:

1) The three wires are separate - not part of a flat cable bonded together by each wires insulation. Twisting the wires holds them in close proximity, thereby making the cable less susceptible to EMI/RFI.

2) The servo cable, whether flat or separate wires, runs in close proximity to (e.g. touching) other cables or metal objects. In this case the twisting helps equalize coupling among the conductors.

Twisting a flat cable (most servos cables are flat) that is not in proximity to other cables or metal objects (usually the case in RC airplanes) accomplishes nothing.

One more thing, coiling the servo lead has the same effect as placing a ferrite torroid on the servo cable... it chokes off common-mode RF currents. This is sometimes a good thing but not always. You might try this if you experience RFI/EMI problems, for example from a gas engine ignition.