Trading Strategies Based On Digital Filters - page 25
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
Good Job man!
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.Clahn04,
Of course @ posting the code. Please find them attached. Only 2 you will find attached are the optimized FTLM and FTLM_KG. I didn't "create" a RFTL or RFTL_KG.......as least not completely. I generated them in the DFG software and copy & pasted the co-efficients to the optimized FTLM and FTLM_KG.
To answer your question, I expected them to be improvements over the un-optimized versions (which is, indeed, the purpose of optimizing.
) while retaining most of the reactivity.
After re-reading (and noticing things in some of the post I didn't notice before), I'm almost 100% sure I did it incorrectly. I just read up to the 14th page and thought I had it and made the same beginners error that you and dvarrin were experiencing when creating the optimized STLM, in that I was just typing half of the "period" of FTLM to into the delay space to generate the RFTL. As a result, there was smoothing but more than acceptable lagging. Still trying to understand how to get this "delay".
I'm sure that the proper approach was explained in earlier post thus you folks have already helped. (I told ya I was slow to really understand it.) Really appreciate everyone's input and perspective. Thank you all of you.
Peace,
F.F.L.
P.S.
Might as well put it out there before both you, I, and anyone else reading forgets to remind you. @ your question about a strategy.
FFL,
I've got a free minute over lunch, so i wanted to respond. I'm not at my home computer, but will try to clarify your "lag" question.
As we have stated above ( no problem to go over it again!), when you create a FATL or SATL, the DF software generates coefficients for you in the code. Lets say that it generates 15 coefficients ( i'd say 13-18 should be a good number for a SATL). That is your period, not the P1 and D1 numbers. So, if i have 15 periods ( i.e. coefficients), and i want to delay my SATL by half that number, i want to have my delay be 15/2, or 7.5. So, somewhere around 7-8 (+-10%) is the delay you should put on the SATL. I would probably go with 7, but that's just me.
Hopefullly it's crystal clear now. Let me know.
cl
Clahn,I was just reviewing the latest posts of the thread.I noticed that you asked about having basically the same cycle in multiple timeframes..that is an exceptional opportunity..like having a 32 cycle in m30,and a 64 in m15..then you use both on both AND..when the cycles are fitted to the price changes(or to the stlm2),you use them,when they start to "misfit",you stop using them..then,you will see,after several cyclic misfits,that the cycles fit again..go for it ..again..;).
I will probably post a pic/several ones
tomorrow,or,in case the opportunity doesn`t arise,during the rest of the week
Yes, please, Simba
at least couple of TFs scrolling back on moments when it fits and don't
just to get feeling
cycles
Yes, please, Simba
at least couple of TFs scrolling back on moments when it fits and don't
just to get feelingLadies and Gentlemen, the thread is attracting the king of "mtfing"
..fxbs..welcome on board 
See the 3 pics,all have the same 2 cycles..an EURJPY M15 cycle of 60 periods..and a M30 EURJPY cycle of 30 periods...on an EJ M30 chart.
First one,"fit",shows clearly 2 different areas were the cycles are fit to the prices and then become unfit,as they are actually,so,these cycles are of no use..yet
.
Second one shows,within the fit area..a sell area defined by both cycles slopes being negative..so,this is an area to be short,look for shorts at retracements,etc.
Third one shows,within the fit area..first a buy zone,or prepare for buy zone if you wish,where one cycle has already changed slope and the other looks "mature" ..then a long only area where the 2 cycles have a positive slope..and an exit and stay out of market area where oone of the cycles has already changed the slope.
Hopefully,within the few next days ,these cycles will "refit" again,so,I will be able to show some real trades done with them.
FFL,
I've got a free minute over lunch, so i wanted to respond. I'm not at my home computer, but will try to clarify your "lag" question.
As we have stated above ( no problem to go over it again!), when you create a FATL or SATL, the DF software generates coefficients for you in the code. Lets say that it generates 15 coefficients ( i'd say 13-18 should be a good number for a SATL). That is your period, not the P1 and D1 numbers. So, if i have 15 periods ( i.e. coefficients), and i want to delay my SATL by half that number, i want to have my delay be 15/2, or 7.5. So, somewhere around 7-8 (+-10%) is the delay you should put on the SATL. I would probably go with 7, but that's just me.
Hopefullly it's crystal clear now. Let me know.
clCL,
Thanks for using your lunchbreak to share, w/ me, food for thought.
I'm not sure whether or not you were just using those figures as an example because I've never generated an SATL w/ so few co-effiecients. This could be where I'm making my mistake. I had another go @ it and my optimized STLM turned out quite well IMHO. It's the FTLM and FTLM_KG that aren't doing as I believe they would/should do if I was doing everything correctly. Perhaps I'm not using the proper peaks. I will try again. Optimized FTLM is WAY too slow when compared to unoptimized and same for FTLM_KG. The confusing thing is there are, in fact, less co-effiecients in the optimized than the standard one's.
Thanks again for the thorough explanation. It's crystal @ being clear. I will keep trying.
FFL,
Definetly keep trying. In my opinion you should be optimizing your digital filters so that you get few coefficients. If you have an SATL that has 40 coefficients, it's going to be slow in terms of processing. You then have to delay it by 20, which will in effect make a useless STLM. What i have found through testing is that the same SATL that gave you 40, if you tweak p1 or d1, you can get it to less than 20 by putting in different peaks and have it "look" very close to the SATL with 40 coefficients. It's basically guess and check. I'm trying the mtf spectrum that i talked about earlier so we'll see how that goes. What i was thinking about doing was this - Get the data from the first week of january, then run the EA with digital filters for the second week of Jan. Record results. Then get the data for the second week, make filters, reoptimize the EA, and run it against the 3 rd week...etc etc. I'm finally getting some free time again, so hopefully i can proceed with that. We shall see.
cl
Simba,
I was looking at your "fit" stuff. That's fascinating. What are you using to create the indicators ( i.e., hopefully this isn't ignorance, but is it RBCI, or perhaps an SATL in a seperate indicator window not on the chart).
Thanks,
cl
Sorry for all the posts... is there a way to "edit"
Nonetheless....i've been messing with the "fits"....i understand it's RBCI, but i think i need a bit of schooling to understand how to properly construct an RBCI....i've read the help stuff...but when i make them, i am getting 400+ coefficients ( or sometimes 200)...etc....
Thanks,
cl