“Setting” the witness point of the string

[Russ C]

It definitely makes a small difference in action and intonation at the nut - with the thicker strings at least.
I’m sure the full bend in the string would occur sometime with playing anyway, but as a tech who’s paid to make guitars play their absolute best I do it before final detail on the nut because it’s how I can be sure the nut is spot on .. and I know a client will feel the action at the nut before strings get “played in”.

[stormin1155]

I do this. It is especially important when doing a setup where I make adjustments to the nut slots. If I don't, I may make the slots too deep, and as the string settles in it may start to buzz. I've not heard the term witness point used before.

[shufflebeat]

If this was a significant issue then it would be worth considering:

If the curve doesn't settle over time then you can set the action to compensate for it.
If it does then the action on my 13/56 strings would noticeably drop over time, it doesn't.

[canyongargon]

This is important in my estimation when doing a setup, especially at the nut. Frank Ford mentions this in his article on setting nut action as step 1. As someone else mentioned a nut slot that seems perfect can end up being too low once the string develops a bend at the point of contact with the nut.

The same type of “bend” at the saddle can affect the overall action, and the same principle of pushing the string down in front of the saddle applies. I do both when changing strings as a matter of habit, but only after the strings are stretched thoroughly to the point of holding steady tune. My experience is that setting the bend in the string before stretching can cause the small bend to ride up into the nut slot as you tune up over time and cause some strange buzzes.