Trading Strategies Based On Digital Filters - page 24

New comment
 

I found a tool over at forexfactory called VHands. Some of you may be familiar with it. It basically is an EA that turns the strategy tester into a forward testing terminal. All quote data is streamed in at varying speeds ( you pick it). You can attach technical indicators and such, and trade as if you are in real time. I have found it helpful for me because as price is fed in, it changes the digital filters accordingly. So, you can see "how" they form and turn and such, which is important for writing an EA. Just thought i would leave that tid bit here.

I tested the no mod on EURJPY this weekend. The results were not good. They are on my other computer so i can post them later. I guess it leads me to a couple of questions. Are my indicators "off". Or, is EURJPY too volatile of a pair. The bare bones slope system may work better with "slower" pairs, i.e. EURUSD as Simba showed. Anyone else try EURJPY or GBPJPY for that matter? Perhaps i did something wrong.

cl

 

It is on this thread https://www.mql5.com/en/forum/general

 
clahn04:
I found a tool over at forexfactory called VHands. Some of you may be familiar with it. It basically is an EA that turns the strategy tester into a forward testing terminal. All quote data is streamed in at varying speeds ( you pick it). You can attach technical indicators and such, and trade as if you are in real time. I have found it helpful for me because as price is fed in, it changes the digital filters accordingly. So, you can see "how" they form and turn and such, which is important for writing an EA. Just thought i would leave that tid bit here.

I tested the no mod on EURJPY this weekend. The results were not good. They are on my other computer so i can post them later. I guess it leads me to a couple of questions. Are my indicators "off". Or, is EURJPY too volatile of a pair. The bare bones slope system may work better with "slower" pairs, i.e. EURUSD as Simba showed. Anyone else try EURJPY or GBPJPY for that matter? Perhaps i did something wrong.

cl

Hi clahn,I presume that you backtested it on Eurjpy by using standard RSTL...MY COMMENTS ARE :

1-Please post your test results,detailed statement,so that we can check several factors,among them:Length of test,timeframe,indicators used,risk taken,etc,etc,etc..nothing like a detailed statement,with the full list of trades and a little explanation,for everybody to be able to learn and comment.

2-It is not proved ,even for EURUSD that no mod works with non custom RSTL,What was tested was the custom RSTL for EURUSD H4 for a very short period,and ,surprisingly,this strategy gave good results.But testing with standard RSTL,for a long enough period,hs not been proven to work,even for

EURUSD..try it..

3-In my opinion,No simple strategy,unless using custom digital filters,and reoptimizing them every week,will work for any pair and timeframe,specially on the JPY pairs..you can just test them for around 3 years and you will see by yourself...if they survive July/August 07,then they will be destroyed in December 07 to now

4-IMHO..First we need to discard several strategies,and to do so,we have to prove either that they didn`t work or that they didn`t work well enough..to do so,we have the EA and the data..so,I would say that if we use the 3 EAs posted,with standard SATLs and RSTLs,for around 3 years of data,for EURJPY,EURUSD,GBPJPY,GBPUSD..AND we post the detailed statements,we can start with the right assumptions.

5-Once we discard the wrong assumptions,we can take 2 ways,simultaneously or sequentially,if enough people want to cooperate:A-Create and test sophisticated strategies with standard digital filters..and B-Reoptimize the custom digital filters,every week ,for around 200 bars and then forward test on next week.

6-There is an additional way,thanks to Jojo,we have an alternative mean to analyze the representative cycles,we can create cyclic indicators based on Goertzel cyclic analysis...there are several problems ,for the moment at least:First,we can know the cycles,but we don`t have the means to translate them into indicators,since the DFM doesn`t work for periods with lengths as high as 661,etc Second:We still do not have the capability of real time adaptive timing,which would be the perfect way to use the filters,self adapting to the latest x dominant cyclic periods,etc..if Jojo can do it,we will probably have to rethink how and what we test,since this would be a huge step ahead.

 

Simba,

I used a custom RSTL based on spectrum analysis, not a standard. For some reason i only got "buy" orders also. I have not had a lot of time to get in depth with all of this as i have been busy at work. Nonetheless, i will post everything here and hopefully you guys can discern what is going on. I may have made a mistake ( not that too uncommon:-)

cl

 

Just a few thoughts.....These are just questions that have come into my head, and i figured if i was thinking about them someone else is too. So, it would be great to keep this running dialogue we have to help everyone learn.

The first question that jumps out at me is this. Suppose every week, you do a spectrum analysis on the last 200-204 bars of tic data to get your coefficients. Would the best approach be to do this analysis over multiple time frames, and then find COMMON peaks throughout the time frames? I, for example, like a 15 minute chart. If i do a 4 hour analysis, my 200 bars are taking me back 33 trading days (roughly) to the end of december ( from today). 200 15 minute bars takes me back to February 14. Since the markets are fractal, one would expect to be able to use either? I would say yes, but i guess i'm just wondering the validity of this.

I was trying to attach 4 spectrum analysis i did (4hr, 1hr, 30min, 15min) but i don't have cropping software on my new computer. I have duel monitors so it screenshots the whole thing and makes it too wide.

In any event, hopefully i made sense. This topic may have been discussed before, but i figured i would ask it anyways.

cl

cl

 
clahn04:
Simba,

I used a custom RSTL based on spectrum analysis, not a standard. For some reason i only got "buy" orders also. I have not had a lot of time to get in depth with all of this as i have been busy at work. Nonetheless, i will post everything here and hopefully you guys can discern what is going on. I may have made a mistake ( not that too uncommon:-)

cl

Your no mod EA was profitable and with a profit factor of 1.55.the code looks right,don`t know why it opens only buy orders..so,where is/was the problem?;)

 

Hi!

I do some search & found an intersting article about Digital Filters and Trading Strategies Based On Digital Filters with PF 17( that published in the "Valyutny Speculyant" magazine in 2000-2002 )

Link

Does any body know how they buy and sell ?

(they said this information and maybe is wrong)

 

The reports i get are off it seems. I guess that was the problem. One time i'll get a test that is negative, one time it's positve. But i guess it poses a question. Why would the one hour be so unsuccesful while the 4 hour is the opposite? Is this because of the "issue" that i asked about a couple of posts ago ( i.e. you should do the analysis on the timeframe you wish to trade?)_

cl

 

Howdy all,

Hope everyone had a good weekend.

I read the thread and applied the methods to create a fitted FTLM and FTLM_KG to GBP/JPY H1. Thanks everyone for the educational discussion that allowed me to do so. Unfortunately, it didn't turn out quite how I expected (pics attached). The optimized FTLM looks more like un-op FTLM_KG and the optimized FTLM_KG lags ALOT when compared w/ the un-optimized FTLM_KG. I know I messed up somewhere along the way. Not only do the optimized indies lag, but they are less smooth. Not sure what I did wrong. Any ideas?

Peace,

F.F.L.

P.S.

I have some ideas for an e.a. using these 3 indicators. Have to figure out this optimization thing first.

 

Forex_for_Life,

Could you please post the indicators so we can see the code, along with the FATL and FSTL so we can see the settings? Just by looking at the pictures, the indicators seem "ok". What were you "expecting"?

Hopefully we can help.

Thanks,

cl

P.S. - Simba alluded to a use of the STLM MTF indicator. Remind me to get into a strategy question about i have that after we get yours answered and look into your FTLM idea.

1...171819202122232425262728293031...138
New comment