If we go back to the days of Howard Hill, he made the comment, probably somewhat in jest, that he didn't believe he was a skillful enough archer to shoot a recurve bow. The reason he said that was because whatever recurve bow he was shooting would put arrows high and low and all over the place without any change in input from him, as far as he could tell. I guess when we get unexplained flyers, we can take some peace in that: if it could happen to Howard Hill, it could happen to anybody.
Howard didn't explain why this was happening to him; it just convinced him he should stick with his longbow. My own theory is that he was probably heeling down on his recurve like he did on his longbow, which caused one of the limbs to bend more than the other. This might have been okay if he could get the same exact bend every time he shot it, but once something like that gets out of balance, it gets very sensitive and could bend more one time than the other without the shooter being able to tell the difference.
If that is true, then probably one of the most important components of a bow to make it forgiving would be to shoot it the way it was designed to be shot. In other words, put grip pressure where it is designed to be put, or if you don't like it that way, then shoot a bow that is designed to have grip pressure where you want to put it. Of course, this applies to all other components of a bow's design as well.