# Triangular moving average (TMA) ... - page 39

62905

wccmcd:
Yes I have both of them

In Hurst's book he shifted two moving averages and then extrapolated them to get a cross and then predicting the prive move. I assume you have read the book too and not sure if you understand him on the little part -- see the picture.

To me, your TMA is just like a ready-to-use shifted MA: shifted, and extrapolated. But, how did Hurst "smoothed" that full span moving average?

wccmcd

Actually, if I remember correctly, they never explained what kind of extrapolation did they use in their analysis. If we could find out that, then it would be easy to do something like that

102

No they didn't. Actually you have done something even better...

I have a strong feeling that by using TMA, I'm approaching the right track... Again, thank you Mladen, for such a great work.

wccmcd Actually, if I remember correctly, they never explained what kind of extrapolation did they use in their analysis. If we could find out that, then it would be easy to do something like that
3960

:) That is the step that most skip, and then they wonder why they can't they trade the TMA. Also, Brian Millard's book (Channels & Cycles: A Tribute to J. M. Hurst) is a very good one and maybe explains even better how to use centered TMA

"Channels & Cycles: A Tribute to J. M. Hurst" Brian Millard :the book

This book is for the technician or trader who wishes to think for himself, and not rely on over-used, and often misapplied mathematical techniques.

Millard instructs you on how to use mathematical tools to extract cycles from stock price data, predict price and time zones, and determine the optimum time period to trade in.

Through simple logic, Millard demonstrates the optimum time period to trade in is between 50 and 100 days. If you are a diehard daytrading scapler, you probably won't like the conclusions he comes to in this book.
676

Guys, thanks for the info and the link.

I was always wondering why mladen always talked about Hurst and Millard when talking about centered TMA

62905

sebastianK:
Guys, thanks for the info and the link. I was always wondering why mladen always talked about Hurst and Millard when talking about centered TMA

sebastianK

Just to clarify : that is not a holly grail. But, once when you read how they intended to use it, it becomes a useful tool for trends and cycles estimation (not for signaling, as I have told a zillion times already ).

102

It's not a holly grail at all, because -- cycle period tends to change from time to time... but the good part is, by using TMA, at least you can have a price moving direction in your mind now.

For now, my problem is to determine WHEN, or at what point, the trend starts to change. Is that since the TMA change its color?

The second is, how to get the cycle period parameter -- what number should I set TMA's halflength to? I know there are some tools to extract the dominant cycle period, but then what? put that number into TMA? not sure if that's going to work, trying now...

sebastianK Just to clarify : that is not a holly grail. But, once when you read how they intended to use it, it becomes a useful tool for trends and cycles estimation (not for signaling, as I have told a zillion times already ).
102

I'm playing with your "regular" TMA (not centered) now. I found it's also interesting. Do you have a good regular TMA, as good as your CenteredTMA, with bands, MTF and shift function? Thank you.

23

Mlanden, hope i'm not intruding --I've only recently joined -- this is a great thread

Hurst also advocated using an - inverse MA " to help with the timing of turns

62905

calisto77:
Mlanden, hope i'm not intruding --I've only recently joined -- this is a great thread Hurst also advocated using an - inverse MA " to help with the timing of turns

calisto77

Do you happen to have the formula of inverse ma?

9195

wccmcd:
Mladen, I'm playing with your "regular" TMA (not centered) now. I found it's also interesting. Do you have a good regular TMA, as good as your CenteredTMA, with bands, MTF and shift function? Thank you.

Wccmcd, you can try this version of regular non recalculating Tma with bands and judge for yourself what is better.

Files: