You are correct that using very long servo arms to get 3D thows gobbles up servo torque. But that is not friction, it's bad mechanical advantage. After learning a hard lesson early this spring I will unlikely use the outer most hole in the servo arm and the inner most hole on the control horns anymore. It may give you monster throws, but it will likely strip your servo unless it is grossly over sized. It's all about trade offs. And the trade off to huge throws is not having as much of the servo's torque working the control surface. If this confuses anyone, I'd say always go 1:1 as far as the distances from the hole to the axis of rotation on both sides of the control rod. That way if you have a 100 oz servo, the control surface will see the entire 100 ounces (less a little bit of friction).
As far as crossing over pull pull cables, it makes one of the cables go slack when you run the control surface to one end or the other. I did this on my CAP232 120 and it really works fine. the slack is not an issue. There are ways to fix this with crossing cables but I really didn't feel like taking it any further (it flies fine).
ORIGINAL: Lancair-RCU
Thanks, thats pretty much what Id figured.
With the power of todays servos, losses to friction will be minimal. We're losing more power by using larger than normal control horns on our servo's. Im not sure where I read it, but you loose a LOT of potential torque the further from the servo centre you get. I endeavouring to use standard servo horn spacing but will probably need to go a size up.