When I first read this thread I was hoping Bryan H would chime in. I knew what his answer would be, but better coming from him than me. I've been flying Bryan's designs exclusively for about 4 years, and I've never once seen any of them require anything other than changes to CG and wing incidence to correct a pull or push on knife-edge (or up and downlines for that matter). In fact, I've seen elevator halves so minutely off yet cause non-stop trouble when pulling an upline. "Twisting" the stab is not the answer.

Going beyond what Hebert addressed, here's a little theoretical stuff on P Factor: P Factor is a potential for issues only when at a very high angle of attack. Even then, the spiral slipstream is irrelevant during knife-edge because even if it is there, the tail is probably below it due to the high angle of attack. The only part of P Factor that could come into play is the uneven lift (thrust) generated by the prop because the downward side is producing more thrust than the upward side. When this happens you will see the airplane pull towards you when flying right to left, and pull away when flying left to right, the direction of rudder used for the knife-edge would be irrelevant to this cause. For the most part, though, we don't fly a high-alpha knife-edge in pattern.