| Welcome Guest, please sign in to participate in a discussion. | Search | Active Topics | |
Registered User
Joined: 10/7/2004
Posts: 816
|
Hi Bruce,
Is it possible to create a realcode zero lag moving average that I could use with a 5 period Linear Regression Slope
Thanks
Bob Mc
|
|
Worden Trainer
Joined: 10/7/2004
Posts: 65,138
|
You could just add a DEMA or Linear Regression to Linear Regression Slope indicator as a child indicator. Both the DEMA and Linear Regression are types of zero lag moving averages.
-Bruce
Personal Criteria Formulas
TC2000 Support Articles
|
|
Registered User
Joined: 10/7/2004
Posts: 816
|
Interesting - who knew? I certainly didn't
I had tried a 3 period exponential MA as my best thought
So, I now have tried all 3 -exponential; DEMA and Linear Regression. They are all very similar and frequently have the same turning point on the example I used. DEMA and Lin Reg have slightly wider ranges and the Lin Reg sometimes turns earlier than the other 2.
Now, i just have to find the "right" Linear Regression Slope to use.
Regards
Bob Mc
|
|
Registered User
Joined: 12/31/2005
Posts: 2,499
|
QUOTE (BobMc)
Interesting - who knew? I certainly didn't
I had tried a 3 period exponential MA as my best thought
So, I now have tried all 3 -exponential; DEMA and Linear Regression. They are all very similar and frequently have the same turning point on the example I used. DEMA and Lin Reg have slightly wider ranges and the Lin Reg sometimes turns earlier than the other 2.
Now, i just have to find the "right" Linear Regression Slope to use.
Regards
Bob Mc
Well if the right slope was obvious it would stop working.
You might consider some additional longer term setup circumstance relating to volume ot atr or some other indicator situation that would improve the perfromance of the selected "obvious" slope. or some overall market circumstance might improve the performance.
Without the additional setup there would be a lot more trades, with the setup the number of trades would decrease but the perfromance would improve. To avoid curve fitting the number of trades should be in the hundreds or even thousands, not reduced to just a few.
|
|
Registered User
Joined: 10/7/2004
Posts: 816
|
Thanks JAS,
I do have several other qualifiers within the layout I'm working with.
I understand about "curve fitting" - even if I'm only working with one stock or index - I can see where it can make a big difference in trending vs cyclical times.
Appreciate your input
Regards
Bob Mc
BTW- I've been "experimenting" with Starc Bands applied to various indicators and there seem to be some good actionable action/reaction possibilities when the indicators reach or exceed the bands in either direction. Hve you had any experience with them?
|
|
Registered User
Joined: 12/31/2005
Posts: 2,499
|
QUOTE (BobMc)
Thanks JAS,
I do have several other qualifiers within the layout I'm working with.
I understand about "curve fitting" - even if I'm only working with one stock or index - I can see where it can make a big difference in trending vs cyclical times.
Appreciate your input
Regards
Bob Mc
BTW- I've been "experimenting" with Starc Bands applied to various indicators and there seem to be some good actionable action/reaction possibilities when the indicators reach or exceed the bands in either direction. Hve you had any experience with them?
No not exactly Starc bands. but I often use normalized moving average ATR differences, (ATRx - ATRy)/C, as a filter. As the Stark upper band is SMA + k*ATR and the lower band is SMA -k*ATR, this is similar to a normalized ATR and is is plotted with price, thus normalized. A contracting ATR would increase the chance of a crossover when there is price action toward the bands.
--------------------------------------------------------------------------------------
A lot of the fertile ground in TA relates to reversion to the mean. The 10 day low being a typical example.
|
|
|
Guest-1 |