I will be very appriciate for any suggestions about my problem. In the world of filtering for financial markets time series that are nongaussian (multifractal) there are very small amount of causal (i.e. do not look into the future) filters that are described in some special literature. But all that from XX century. For many years I've used Jurik moving average (for MATLAB version): http://www.jurikres.com/faq/faq_ama.htm#betterthan. And I don't believe that there is no filter in cutting edge digital signal processing science better than JMA.
So my question is: Do you know causal adaptive low lag filter better than Jurik MA?