Trading Strategies Based On Digital Filters - page 21

SIMBA
2078
SIMBA  
deeforex:
Thanks for the EA Simba.

I just kept everything as default, with the exception of changing the names of the indicator. I loaded it with the indies that I created last week.

I've got 1 order on it so we'll see how it does. Since I took the pic, the order is down 10 pips to +5pips.

dee

deeforex:Thanks a lot,I presume what I mention in point 2 is what you did,please be so kind to confirm

1-This is not an EA for trading..it is an EA for testing strategies,so,if you want to test the rstl standard change of slope strategy,better to do it on enough historical data..suppose you make money the first trade,so what?Statistically you need several years of testing to support or disprove the concept.

2-The interesting strategy to test,historically since 12 of December and forward live for this week,would be to use the customized SATL and RSTL,we created from short term optimization,I presume this is what you did,can you confirm?...since I am not at my pc,I don`t have them...so,I would like you or anybody else,to please substitute them in the EA and test them Live...if you don`t mind

3-The idea,at least my idea,is to optimize on past 200 bars of data,then test live(with demo) next week..reoptimize at the end of week and then retest again the next week again,etc,etc...AND to compare several different entries and exits from the strategy menu..NOT to just put an EA with standard indicators and see what happens,because we are not going to learn anything from it,if it makes money or not in the first 2 or 3 trades(that would take around a week per trade) is statistically irrelevant,so ,after 3 weeks we won`t know anything new about it.

4-Basically,I don`t intend to discover the Holy Grail,my aim is just for everybody to use the following mindset:Discover optimized settings or customized digital filters,test them with the test EA I provided to show which of the trading strategies shows more promise,use this trading strategy forward test(on demo)with frequent reoptimizations...and learn..so,that when we think we have discovered the Holy Grail,before jumping headfirst,at least we cool down a little by some reality check,which only testing with an Ea ,or several years of experience,can provide...

Of course this is just my aim,so,anybody do what you wish

SIMBA
2078
SIMBA  
clahn04:
Forex For Life,

That's funny. :-) I promise i didn't steal it! lol

Simba,

Thanks for the response. Last night i did something similar in excel, to test the SATL slope. I found that filtering out the slope when it's close to level and not moving much kept you out of a lot of choppy trades. I think for my purposes i found the 2 period slope at each point in time. I then averaged the past 2 periods of that. If it was >.025, we have a Long trade. If it's < -.025,we have a short. Anything in between represented "No trade". Attached is the word file ( i can't post excel files, so i pasted it into word). The only thing i'm wondering is when the entry should be. Should it be at the beginning of the bar, so the slope is referencing the previous SATL for slope or not......not too sure. The SATL will always recalculate with new data....

I think there is some great power to be found in these slope strategies.

Comments always welcomed. I'll be back around later.

Safe trades,

cl

clahn,entry should be at open of next bar..same for exit or exit and reverse,the way it is coded in the EA for example,once rstl is higher than previous rstl(both calculated at close)it opens a buy order at the open of the next bar..you can check visually with the strategy tester..anything else would be lying to ourselves.

I believe it is possible to code the minimum required slope into the EA,just by adding a new ext internal "slopefactor=x" and changing the code of buy and sell triggers BUY 1_1>(BUY 1_2+SLOPEFACTOR)..FOR BUY 1_1 being actual SATL and BUY 1_2 previous one,this is much better for testing since we can check for several alternatives,currency pairs,etc...will try to modify the posted EA to be able to do so.

SIMBA
2078
SIMBA  

EA with slope

The EA is coded to buy or sell by change of slope in the satl of a minimum defined by slopefactor,coded at 0.25,but easily testable and modifiable..exit has been coded by just a change of slope of the satl from positive to negative,etc,obviously,you can modify it to suit your (testing) needs

deeforex
93
deeforex  
SIMBA:
deeforex:Thanks a lot,I presume what I mention in point 2 is what you did,please be so kind to confirm

was your point 2 this...

And you leave this code as it is

if (Buy1_3 > Buy1_4 ) Order = SIGNAL_BUY;

if (Sell1_3 < Sell1_4 ) Order = SIGNAL_SELL;

if (CloseSell1_3 > CloseSell1_4 ) Order = SIGNAL_CLOSESELL;

if (CloseBuy1_3 < CloseBuy1_4 ) Order = SIGNAL_CLOSEBUY;

if so, yes, I left this the same.

Thanks for your clarification of your intentions for sharing the EA. I threw the EA on for forward testing just to see how it would do with the "fresh" indies that I created last week. I will work on the aspect of optimizing this coming week.

Linuxser
5921
Linuxser  
moha00:
Hi !

I am a newbia at Digital Filters .

What I must read ? can any body help me?(for knowing them)

Thanks

By reading this section plus some material of John Ehlers (just do some search in the forum) and books like Mesa and Trding Markets Cycles, Rocket Science For Traders and Cybernetic Analysis for Stock and Futures you should be fine.

www.mesasoftware.com is good place to start also.

Mohammad Rahimi
89
Mohammad Rahimi  

Hi !

I am a newbia on Digital Filters .

What I must read ? can any body help me?(for knowing them)

Thanks

clahn04
150
clahn04  

I would add Hurst to that list...

Sorry i've been MIA...i've been crazy busy at work... i hope to be back contributing to this board tonight....i wanted to throw a question out there thought right now....it somewhat pertains to this line of thinking...

The new Futures magazine (Feb 08 edition) has an article in it by Matt Reynolds entitled "Pinpointing entries accross time frames". In it he gives the basic talk about confirming on MTframes an entry. Now i'm not talking about a MTF stochastic strategy or anything like that, but for our purposes, would something like this make sense?..

Example - RSTL on 1 Hour chart is a postive slope, signaling an uptrend. Go down to a 15 minute chart for your entry and exit, and have another qualifier - perhaps the RSTL on the 15 is positive, or SATL ( which is less delayed). Obviously the essense of digital filters is filtering out noise, making our signals more "accurate". I'm wondering if MTF would be disadvantageous because the RSTL basically tells you the direction already. Anyways, just a thought. Feel free to comment.

We have a great discussion going here. Let's keep it going.

Safe trades,

cl

clahn04
150
clahn04  
SIMBA:
deeforex:Thanks a lot,I presume what I mention in point 2 is what you did,please be so kind to confirm

1-This is not an EA for trading..it is an EA for testing strategies,so,if you want to test the rstl standard change of slope strategy,better to do it on enough historical data..suppose you make money the first trade,so what?Statistically you need several years of testing to support or disprove the concept.

2-The interesting strategy to test,historically since 12 of December and forward live for this week,would be to use the customized SATL and RSTL,we created from short term optimization,I presume this is what you did,can you confirm?...since I am not at my pc,I don`t have them...so,I would like you or anybody else,to please substitute them in the EA and test them Live...if you don`t mind

3-The idea,at least my idea,is to optimize on past 200 bars of data,then test live(with demo) next week..reoptimize at the end of week and then retest again the next week again,etc,etc...AND to compare several different entries and exits from the strategy menu..NOT to just put an EA with standard indicators and see what happens,because we are not going to learn anything from it,if it makes money or not in the first 2 or 3 trades(that would take around a week per trade) is statistically irrelevant,so ,after 3 weeks we won`t know anything new about it.

4-Basically,I don`t intend to discover the Holy Grail,my aim is just for everybody to use the following mindset:Discover optimized settings or customized digital filters,test them with the test EA I provided to show which of the trading strategies shows more promise,use this trading strategy forward test(on demo)with frequent reoptimizations...and learn..so,that when we think we have discovered the Holy Grail,before jumping headfirst,at least we cool down a little by some reality check,which only testing with an Ea ,or several years of experience,can provide...

Of course this is just my aim,so,anybody do what you wish

Simba,

I completley agree with you. It may be of some use to discuss as a group what types of strategies we should test (i.e., Slope SATL change, SATL slope beyond a certain factor change, RSTL, etc). It may feel a bit like school, but if we break it up so that as a group we are testing multiple currencies during the week ( each person testing one), we may begin to gain some ground. Just my $.02.

cl

forex_for_life
368
forex_for_life  
clahn04:
Simba,

I completley agree with you. It may be of some use to discuss as a group what types of strategies we should test (i.e., Slope SATL change, SATL slope beyond a certain factor change, RSTL, etc). It may feel a bit like school, but if we break it up so that as a group we are testing multiple currencies during the week ( each person testing one), we may begin to gain some ground. Just my $.02.

cl

Yo clahn and other interested parties,

You may have seen this already. It's a smoothed version of FTLM called FTLM_KG (the KG being the initials of the coder who added the smoothing). I use it is to give me the short-term cycle since traditional FTLM is too "leading" for my taste. It, of course, exhibits added lag which is the consequence for smoothing, but it is still damn quick. Screenshot provided for initial visual comparision. This is before any optimizing. Don't know how much room for improvement and I'm still learning how to do it w/ FTLM/STLM/RBCI-type digital filters. I digress.......enjoy.

Peace,

F.F.L.

Files:
deeforex
93
deeforex  
forex_for_life:
You may have seen this already. It's a smoothed version of FTLM called FTLM_KG (the KG being the initials of the coder who added the smoothing).

FFL,

Thanks for posting. Glad to see someone else joining us. Question for you. The 2 indies that you used for your screenshot, do they have the same coefficients inside the code?

the part that reads

value1 =

0.216679747671846*Close

+0.211269420463811*Close

...

-0.000413583883278278*Close;[/CODE]

value2=

0.0364298167208155*Close

+0.0553906231378775*Close

...

-0.000218687638971784*Close;

[CODE]

...etc

value3....

value4...

The reason why I ask is because if these values are not the same, that could be the cause of the lag. I checked out the code and besides the coefficients being different from 1 that I have, another noted difference is the number of CountBars used and I don't think that should impact the results. I could be wrong but I changed the bar count and that doesn't seem to change the results.

Thanks for sharing tho.