Trading Strategies Based On Digital Filters - page 16
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
How many periods are in a 3 period SMA?Yeah 3..well the formula just takes (close2+close1+close0)*0.33333333...so,it is applying a fixed 0.333333 coefficient to 3 periods..Now...How many periods do we have in a Satl that is applying variable coefficients to periods from 0 to 64? And how many periods do we have for a RSTL based on variable coefficients applied to periods from 0 to 90? Ok,now,if you want to do a RSTL for the TestSatl that you calculated,what would be the acceptable range(tip:+-10% on theoretical value) of periods for a TestRSTL ?
I think for SATL then we have 65 periods and for RSTL, I often could see around 90 or 98 coefficients, so meaning that it should be a period of 90 or 98, right?
So for the creation of RSTL, we should take half more than what we have for SATL?
I tried to modify P1 and D1 in DFM, and could see that it is modifying the number of coefficients used, but still I don't understand what they do and don't understand when one is better to be changed rather than the other one.
I think for SATL then we have 65 periods and for RSTL, I often could see around 90 or 98 coefficients, so meaning that it should be a period of 90 or 98, right?
So for the creation of RSTL, we should take half more than what we have for SATL?
I tried to modify P1 and D1 in DFM, and could see that it is modifying the number of coefficients used, but still I don't understand what they do and don't understand when one is better to be changed rather than the other one.1-Yes,right on all counts..65 and 91..and yes there is another RSTL with 98 periods..basically if satl has 65 periods RSTL should have 65+(half less one)..with a margin of +-10%-..so,as a rule of thumb anything between 89 and 108 would be ok..usually ,it is better to use the shorter periods,signal is still smooth and slightly faster
2-P1 and D1 select the cutoff frequencies(in reality they select the cut out periods)...basically you are saying that anything OUTSIDE these periods is of no interest to model a market and timeframe,so,you have to be very careful with what you select.Basically you are assuming the paradigm that the cycles with "not enough height" are probably noise,and the rest are what you will use to model a market.
Why don`t you try to do a RSTL for your testSATL?And then we can proceed to go for a custom STLM2?
1-Yes,right on all counts..65 and 91..and yes there is another RSTL with 98 periods..basically if satl has 65 periods RSTL should have 65+(half less one)..with a margin of +-10%-..so,as a rule of thumb anything between 89 and 108 would be ok..usually ,it is better to use the shorter periods,signal is still smooth and slightly faster
2-P1 and D1 select the cutoff frequencies(in reality they select the cut out periods)...basically you are saying that anything OUTSIDE these periods is of no interest to model a market and timeframe,so,you have to be very careful with what you select.Basically you are assuming the paradigm that the cycles with "not enough height" are probably noise,and the rest are what you will use to model a market.
Why don`t you try to do a RSTL for your testSATL?And then we can proceed to go for a custom STLM2?Thank you for your answer Simba!
I'll create the RSTL, but I need to ask you one more question before doing :-) does it mean that to create a SATL of period 63, I should take P1 a little bigger than 63 and D1 a little smaller (to filter the cycles with a period close to 63), and make sure that the number of coefficients will be 63? Or should I choose the values of the two highest peaks and use them as P1 and D1?
How to ,another option
Dvarrin,you want to trade EURUSD H4 and learn about digital filters,so,let`s say that I have the same interests as you and have created some custom SATL for that pair and tf..This is what I would suggest that you do
1-Check the pic and the frequencies
2-Check the Satl I created and how they relate to 1
3-Try to do,trough trial and error,an RSTL for this SATL
Dvarrin,you want to trade EURUSD H4 and learn about digital filters,so,let`s say that I have the same interests as you and have created some custom SATL for that pair and tf..This is what I would suggest that you do
1-Check the pic and the frequencies
2-Check the Satl I created and how they relate to 1
3-Try to do,trough trial and error,an RSTL for this SATL1. So I notice that there are 3 main peaks at 14, 34 and 115
2. You took 115 for P1 and 14 for D1, and there are 16 coefficients for SATL. So it means that the period is 16? (I'm really lost)
3. Then for RSTL I took 115 for P1 and 34 instead of 14 for D1, because it makes an error in the code with 14 (0*Close...)
Then I simply changed the delay value so that the curve is changing direction when the price is reaching a top or bottom. It looks quite close to another RSTL I've downloaded :-), but there is no logic or rule for what I did and also, as we said before, the period for RSTL should be one and half bigger than the one for SATL, so if SATL has 16 coefficients, I should have only 24 for RSTL, but I've a lot more.
Ah i see....
Well this leads me to what will hopefully be an easy analysis question,...
When doing the spectrum analysis, what is the proper number of bars to use?
I had been using the full history ( so 2000+)....is their a rule of thumb to look for when deciding how many bars to use...
thanks,
cl
2 pics
SATLs are based on close,so use it (the .mq4 at my previous post)with a line chart..preferably
.See the 2 pics,a very simple way to trade this Satl..if close is above satl, then buy..if close is below, then sell..variation 1..add AND if SATL slope is up or down for above and below,later but safer entries...variation2..just go for the slope or the above/below,whichever comes first,we are paid to take risks,so,take them 
...Depending on your propensity for risk,you now have 3 options to choose..choose the one that feels right for you.
Btw ,markets are fractal in nature..try this(devised on H4 tf,and a month and a half specific history ) SatlEurusdH4 on a monthly chart of EURUSD,from 1990 until now,for example,,you will be surprised...and,if you like it,post the pic for everybody to meditate on
dvarrin...
the error in the code is that there is no operator in front of the 0*close.....just put a + sign in front of it.
cl
Ah i see....
Well this leads me to what will hopefully be an easy analysis question,...
When doing the spectrum analysis, what is the proper number of bars to use?
I had been using the full history ( so 2000+)....is their a rule of thumb to look for when deciding how many bars to use...
thanks,
clYes...wanting certainty in a world of chaos is a trait of human nature.But certainty means that your competitors too are certain,so,no market to be traded,everybody wants long,nobody wants short..Price skyrockets ..for the trader that bought yesterday,not today
There are 3 "useful" ways..1-All bars in history..2-All bars since market changed paradigm..3-the minimum that the software allows..The uncertain choice is yours..In my opinion,2 and 3 are the way to go,but this is just an opinion.
And,indeed,chaos is "almost" deterministic..
Simba,
I think you have helped us to realize something truly special...i'm afriad i won't get anything done this weekend now lol..
cl