Mladen, thanks a million times.
Interesting expresion "imitation of approximation"
Actually, I was also thinking of adding internal smoothing
in the line of code which represents %FastK of MACD.
In pure stochastic oscillator this is done with summation but as I see
Doug Schaff omitted it.
%K = 100*(MACD - MIN (LOWMACD, STCPeriod) / MAX (HIGHMACD, STCPeriod) - MIN (LOWMACD, STCPeriod)
%K = 100*SUM (CLOSE - MIN (LOW, %KPeriods), Sk) / SUM (MAX (HIGH, %KPeriods) - MIN (LOW, %KPeriods)), Sk)
Sk means %K Slowing Period
I am again and again amazed how can such tweaks improve the originals.
For example your correction of T3 outperforms the "original" T3.
Gentlelmans,I found samthing similar to my idea and this is it: