Trading Strategies Based On Digital Filters - page 64
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
cycles
K,
2-The cycles:
Test results with the EAs are not biased..Let`s presume H1 Timeframe...if for example you entry long when cycle slope turns up and exit and reverse when cycle slope turns down..Imagine the following case bar1(previous bar)has closed and the cycles turned up at bar close..so at open of actual bar,bar0, EA enters a long-Like you will do if in front of the screen at that moment-then after 1 hour,at close of bar0,the Cycle Repaints down like crazy again..then the EA closes the long,at a loss and opens a short-like you would if in front of the screen at that moment-...then the new slope continues for 10 bars,so the EA just keep the shorts open..then when it turns up again,at bar close,next open the EA closes the short(at a big profit)
and opens a long...it continues for 5 bars...etc,etc...So the repainting effect,if any,is included in the results,it punishes the system,exactly like it would punish you if you where trading it manually in front of the screen....There is no future leak,at close of bar1(previous bar)if the slope is up EA enters a long...next close if slope has repainted down,it closes long(suffering a loss) and opens a short...where is the future leak...?When prediction doesn`t match future direction of prices you get punished with losses,when prediction matches the future direction you are rewarded with profits...usually the repainting is minimal,that is why the EA makes profits,it tracks the RIGHT Cycles to track and discards the unstable ones...we have several versions,one of them uses HMA smoothing and the results are practically the same,another works on raw prices and very similar results too..so,we concluded that it is not the smoothing,but the Cycles ...If you use unstable Cycles the repainting will kill your account no matter how much smooth you add...if you use stable Cycles you will make money,and the smooth will just add minimal differences.
The smooth only affects when using high frequency cycles(like 10 periods,etc)...and this means that those cycles are not reliable for long term trading(we tested this),so,which ones are?
Those that work out of sample
...So,test,test,test...to find stable cycles...That is what we use the EAs for.
I just read this and it makes sense however is quite complicated.
regarding this
Those that work out of sample
...So,test,test,test...to find stable cycles...That is what we use the EAs for.
Obviously only the method which you point here is out of sample test to find
the stable cycles. How many measurements you made with the EAs ??
Do you have a statistical database for this and do you consider in this analysis the events like NFP or rate decisions. As those are triggering quite often of a new cycles.
Than it seems that the key here can be kind of 'cycle stability' indicator
not so much S/N. I just wonder how to measure this.
I include those screenshots from 1m. 1m seems to be very unstable, petty the best money is there.
Krzysztof
I just read this and it makes sense however is quite complicated.
regarding this
Obviously only the method which you point here is out of sample test to find
the stable cycles. How many measurements you made with the EAs ??
Do you have a statistical database for this and do you consider in this analysis the events like NFP or rate decisions. As those are triggering quite often of a new cycles.
Than it seems that the key here can be kind of 'cycle stability' indicator
not so much S/N. I just wonder how to measure this.
I include those screenshots from 1m. 1m seems to be very unstable, petty the best money is there...
KrzysztofNo,the best money is not going to be made on M1,at least with cycles.
Statistical database?We test and optimize in sample for the past month,then check out of sample which ones,of the best results,would have worked for the past 18 months(31st December backwards)..we discard a lot of cycles,pairs and timeframes...the few ones we keep,we have found to be,usually,very tradable....EACH WEEK.
We don`t care about NFP,etc,we don`t filter these events out,they belong to the price series,so,why filter them out?
Cycle stability indicator...we don`t use it(we decide based on tests out of sample as I explained above),but Bartels test seems to be the one to use for this purpose..My personal experience is that it is worth nothing,but I can be mistaken.
Nice pictures you posted
...I hope you enjoy the indicator we sent you for free 
Can you post something of value,besides pics of our old indicators on m1?
Pity you don`t know how to trade it...we gave you a steel sword in the Bronze Age,but it is the swordsman not the sword.
Ciao,have a nice week end.
Simba
No,the best money is not going to be made on M1,at least with cycles.
Statistical database?We test and optimize in sample for the past month,then check out of sample which ones,of the best results,would have worked for the past 18 months(31st December backwards)..we discard a lot of cycles,pairs and timeframes...the few ones we keep,we have found to be,usually,very tradable....EACH WEEK.
We don`t care about NFP,etc,we don`t filter these events out,they belong to the price series,so,why filter them out?
Cycle stability indicator...we don`t use it(we decide based on tests out of sample as I explained above),but Bartels test seems to be the one to use for this purpose..My personal experience is that it is worth nothing,but I can be mistaken.
Nice pictures you posted
...I hope you enjoy the indicator we sent you for free 
Can you post something of value,besides pics of our old indicators on m1?
Pity you don`t know how to trade it...we gave you a steel sword in the Bronze Age,but it is the swordsman not the sword.
Ciao,have a nice week end.
SimbaNo problem, I will post Sunday evenning a few cycles for H4 TF for lets say GBPJPY, EURUSD and GBPUSD than you can post yours for the same pair and TF using you methodology and we will compare performance on Monday and Tuesday. OK ??
Regarding sword. I have 23 similar swords in my arsenal (FFT based) with different window shape, different filters etc and similar funcionality and not repainting and I dont think there is a big difference between your and mine
swords.
Have a nice weekend also.
Krzysztof
No problem, I will post Sunday evenning a few cycles for H4 TF for lets say GBPJPY, EURUSD and GBPUSD than you can post yours for the same pair and TF using you methodology and we will compare performance on Monday and Tuesday. OK ??
Regarding sword. I have 23 similar swords in my arsenal (FFT based) with different window shape, different filters etc and similar funcionality and not repainting and I dont think there is a big difference between your and mine
swords.
Have a nice weekend also.
KrzysztofGood to know that at least you show some of your weapons...
No,you will put the prediction,or what your filters are indicating for the pair of your choice on Sunday evening,so that we are on the same foot so to say
Then On Tuesday Evening/Wednesday we can start comparing...I don`t think it is necessary to compare on the same pairs,though at least one of them should coincide,since pair selection is an important part of the method.
Simba
I hope Richcap didn't give up with us as it would be very big lose for this forum.
KrzysztofI'm lurking
I'm lurking
ROFL! That one caught me off-guard
And here is the R-FTLM-STLM-Adaptive indicator built on the previous one (same parameters)
So I was going back on some former post because I finally have a day off (3 days actually) and I JUST realized you posted this adaptive FTLM/STLM indy. I've been in search of something like this for a little while now! The adaptive RBCI indy is nothing to scoff @ either. Anybody who doesn't recognize the value of what you've shared would have to be completely inept in the area of digital filters!! Thanks again, bro!
Now time to run some test. I will report back w/ any findings.
Hi all,
while I was lurking I played a bit with Goertzel.
The pictures below should clarify that, with stationary (or quasi-stationary) AWGN-noised cyclic signal, both MESA and Goertzel give robust and very similar results.
Enjoy ;-)
Hi all,
while I was lurking I played a bit with Goertzel.
The pictures below should clarify that, with stationary (or quasi-stationary) AWGN-noised cyclic signal, both MESA and Goertzel give robust and very similar results.
Enjoy ;-)Richcap,
Interesting (meaning I don't understand most of it LOL!) post, to say the least.
I will preface what I'm about to say w/ this: Once again, I am not, in any way, shape, or form, an authority on DSP, and am asking from the position of a pupil and NOT that of a cynic or critic.
I noticed you said that they both give similar results w/ STATIONARY (or quasi-stationary (???
)) Additive White Gaussian Noised cyclic signal (another ??? 
)
If the markets are considered to be non-stationary, how can we apply the above information? From my limited knowledge, Goertzel is supposed to be ideal for frequency analyzing in extremely noisy data.
Based off my research (from Wikipedia so take it w/ a grain of salt) the reason for using AWGN is that, and I quote, "while.....the model does not account for the phenomena of fading, frequency selectivity, interference, nonlinearity or dispersion.....it produces simple and tractable mathematical models which are useful for gaining insight into the underlying behavior of a system before these other phenomena are considered."
Can you explain to me, in laymens terms, why one would go this route instead of testing the preferred actual data sets initially. Once again, I'm not asking to be sarcastic/condescending/patronizing. Just trying to grasp as much of this as I can.
Thanks in advance.
(Maybe I answered my own question?) Haaaaaaa!!!!!!
Richcap,
Interesting (meaning I don't understand most of it LOL!) post, to say the least.
I will preface what I'm about to say w/ this: Once again, I am not, in any way, shape, or form, an authority on DSP, and am asking from the position of a pupil and NOT that of a cynic or critic.
I noticed you said that they both give similar results w/ STATIONARY (or quasi-stationary (???
)) Additive White Gaussian Noised cyclic signal (another ??? 
)
If the markets are considered to be non-stationary, how can we apply the above information?
Based off my research (from Wikipedia so take it w/ a grain of salt) the reason for using AWGN is that, and I quote, "while.....the model does not account for the phenomena of fading, frequency selectivity, interference, nonlinearity or dispersion.....it produces simple and tractable mathematical models which are useful for gaining insight into the underlying behavior of a system before these other phenomena are considered."
Maybe I answered my own question. Haaaaaaa!!!!!!You are right.
What I really mean is that if we (mainly Krzysztof) want to tell MESA from Goertzl (from SSA), the test signals we have dealed so far are useless.
We should analyze non-stationary, non AWGN added, 'test' timeseries.
Maybe a sequence of known multi tone "tunes" with some sort of multiplicative noise.
We should try to add noise with our experience as traders. I mean, if in a timeseries synthetized from intermittent sine/cosine waves of various frequencies, we see a human readable support/resistance, we would expect the timeseries to move in S/R-aware way.
A S/R cannot change a strong trend but it introduces noise in it.
We should try to model this behaviour somehow...
OK stop dreaming, I return back to my lurking...
Do indicators need generating ?
This is a great thread guys.
I just wanted to clarify one important thing - do the indicators such as RSTL, RFTL, etc.. need to be generated for the currency pair/TF in general or is it a general version for all ? My understand is much better results if generated for the currency pair using the software and spectometer ? Would be grateful for clarification. Thanks !