Quantitative finance formular are mostly based on martingales, Poisson jump, GBM, CEV, etc.. The logic behind it is that martingale means the future could not be predicted, or, EMH (Efficient-market hypothesis).
fair enough. however, EMH is only a hypothesis now, the other side of the story, that the market is inefficient, could also stands -- provided the inefficiency is always found and exploited then disappeared.
so, is there some theory / math model, that is based on non-martingale analysis?