The Hans-Werner Blasel Estimation is a algorythm trying to simulate the best fit of a time-series. This kind of filter have a reduced lag coefficient compared with the original variable, however large period value can generate high distortion. Also please note the original algorythm is actually really really complex, i tried to remake it the best i can.
I don't have the autorisation of the autor to show the original script, i apologize for that
Somes exemples of the indicator
A exemple of how large period input can damage the estimation
Also use this indicator on heikin-hashi chart,it fit better than in a standard candles chart.
There ares some exemples using this estimation as input or others indicators :