Posting a new CenteredTMA version by Mladen.
Quite a few years ago, i was asked to check the ideas explained in the Brian Millard's "Channels & Cycles: A Tribute to J. M. Hurst" book about, what turned out to be, a centered Triangular Moving Average (Centered TMA), and when I did so and wrote the Centered TMA, as if the whole hell broke through and ever since I am forced to tell that I have regretted that I have ever made that indicator
About "centering" as defined by Brian Millard and J.M.Hurst :
the data is shifted by half a period to the past in order to fit the data the best possible to values
that leaves a "gap" at the most current half period bars, and that part must be added using some kind of extrapolation
TMA calculation has a structure that is suitable to use one specific way to extrapolate (not described anywhere else since it is used only in the centered TMA)
as any other extrapolations, that extrapolation is not 100% accurate either
extrapolations are attempt to "predict" the future, and since we all (I hope) know by now that it is impossible with 100% accuracy, that "prediction" will be changed as the new data is coming in
and that is not repainting - that is simple recalculation (since the extrapolation can not be done differently) - not going to repeat the usual mantra what repainting is but :
repainting is a coding error (no coding error in the above)
repainting is a deliberate misleading of people that think that there is a holly grail in order to sell the "wonder" as a final solution (and, I for one, had been repeating from day one that Centered TMA will recalculate the last half period bars) so it does not repaint from that point of view either
one might even find the origin for 'wonder" when one guy took the indicator from then existing forex TSD elite section and promoted it as a wonder system on this same forum, but then, that is a matter of history I guess
About some versions floating around :
apart from the above stuff mentioned, some are still using stuff that was messed up or created years and years ago
for the sake of (my mind) I have made this version that calculates centered TMA with asymmetric bands and checks the execution time (in microseconds, not milliseconds)
average execution time on a new tick (including the recalculation - ie: not just one bar) is from 11 to 16-17 microseconds
PS: not posting the source code. I am not going to fall into the same trap when they are using the code of this indicator to sell yet another "wonder" indicator by simply using the free code and changing a few simple things in it. It is fully operational, multi time frame version with only one "issue" : it will always display that text in the upper left corner - like the example.
Have a nice weekend guys - hope this clarifies the "Centered TMA wonder"
- Centered TMA (asymmetric bands).ex4
- (20.57 KiB) Downloaded 340 times