How often are theories in "opposition to the mainstream consensus among those who are qualified to evaluate its accuracy, such as scientists or historians," valid?