I'm building my first switchbarrel now. I guess I have an overly-simplistic view of torque - my plan is to start at some small figure, fire a string, snug it up some more, fire, etc. until further tightening shows no change.
Why guess when you can collect your own data?