Optimized Trend Trading | Page 10 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Optimized Risk vs Reward Equation 335 replies

Loss Less Optimized Strategy 63 replies

Optimized Parameters in Metatrader Testing? 3 replies

"Optimized Parameter" feature in EA? 0 replies

Strategy with Optimized parameter 1 reply

Attachments: Optimized Trend Trading
Exit Attachments

Optimized Trend Trading

  • #181
  • Nov 23, 2008 2:42pm Nov 23, 2008 2:42pm
AugustLeo I am impressed with your presentations. However, I need some guidance on how to trade the EUR/USD pair. I am new to Forex trading. Can you please assist?
  • #182
  • Nov 23, 2008 2:58pm Nov 23, 2008 2:58pm
Quoting Trendylines
AugustLeo I am impressed with your presentations. However, I need some guidance on how to trade the EUR/USD pair. I am new to Forex trading. Can you please assist?
Please DO NOT pollute this thread with extraneous matters. It is for trend trading and signal processing discussions ONLY. PM folk for private questions!

CB
  • #183
  • Nov 23, 2008 5:12pm Nov 23, 2008 5:12pm
Quoting Trendylines
AugustLeo I am impressed with your presentations. However, I need some guidance on how to trade the EUR/USD pair. I am new to Forex trading. Can you please assist?
Trendylines - I am unable to give you the guidance you seek. Please search the forum to find what you're looking for.

As Codebreaker posted, this thread has a very focused perspective dealing with trend trading and the signal processing that facilitates trend trading. There's a lot of meat to that subject and I hope this thread delves deeply into it.

Thanks.

AugustLeo
  • #184
  • Nov 23, 2008 5:27pm Nov 23, 2008 5:27pm
Quoting AugustLeo
CB - I'm wondering how well your clone will replicate the JMA test rusults on the website.
CB - If you're so inclined, give Jurik's test data a go and show us how well your clone repilicates Jurik's results. I'm curious as to whether you've actually cloned the famous Jurik JMA.

That said, I thank you for all of your lengthy technical posts - very enlightening. I'm enjoying our technical exhange immensely.

Thanks.

AugustLeo
  • #185
  • Nov 23, 2008 6:29pm Nov 23, 2008 6:29pm
Quoting AugustLeo
I thank you for all of your lengthy technical posts - very enlightening. I'm enjoying our technical exhange immensely.

Thanks.

AugustLeo
Me too! It is great that some serious folk are getting involved with the forum. I am very much hoping that between us, we can develop some seriously profitable systems.

I'll have one of my students look at the Jurik stuff - would make a good Final Year or even MSc project. Will report back in due course (could take a while... You know students!)

CB
  • #186
  • Nov 23, 2008 6:39pm Nov 23, 2008 6:39pm
Quoting Codebreaker
Me too! It is great that some serious folk are getting involved with the forum. I am very much hoping that between us, we can develop some seriously profitable systems.

I'll have one of my students look at the Jurik stuff - would make a good Final Year or even MSc project. Will report back in due course (could take a while... You know students!)

CB
I wish that I had had the choice of such a juicy subject for my MS, with an advisor such as yourself.

AugustLeo

PS: I'll be on travel for the next week, so please understand that if I don't reply it's only because i can't.
  • #187
  • Nov 23, 2008 6:44pm Nov 23, 2008 6:44pm
Quoting Codebreaker
Me too! It is great that some serious folk are getting involved with the forum. I am very much hoping that between us, we can develop some seriously profitable systems.CB
i have the technical skills , but must admit that I'm partially uninformed about what is needed to develop a profitable system.

That said, looking forward to plumbing the depths of the financial noosphere ...

Augustleo
  • #188
  • Nov 28, 2008 9:04am Nov 28, 2008 9:04am
Hi Codebreaker,

How do you scale the informational efficiency in a timeseries of financial data?There are some techniques that I read like mutual information,correlation dimension,hurst exponent etc.. how do you implement them in real trading?

There are codes available for Hurst exponent(or fractal dimension).One of them is Sevcik's one.The code is not same thing (it doent measure it right) what it says in the article : http://www.complexity.org.au/ci/vol05/sevcik/sevcik.html we havent' been able to find the''right'' code.
Also what must the sample size be? İs there any standart?

Thanks in advance.
  • #189
  • Nov 28, 2008 12:23pm Nov 28, 2008 12:23pm
Quoting mystified
Hi Codebreaker,

How do you scale the informational efficiency in a timeseries of financial data?There are some techniques that I read like mutual information,correlation dimension,hurst exponent etc.. how do you implement them in real trading?

There are codes available for Hurst exponent(or fractal dimension).One of them is Sevcik's one.The code is not same thing (it doent measure it right) what it says in the article : http://www.complexity.org.au/ci/vol05/sevcik/sevcik.html...
It depends what you want to do. Sevchik's method for measuring fractal dimension is technically (mathematically) very accurate but requires a huge amount (theoretically infinite amount) of data so is, in practice, not very useful.

If you want to measure "Mutual information" then remember that this a technique for comparing two or more data sets. If you have one set, you might look at the "self information", perhaps most usefully in terms of Signal-to-Noise ratio (SNR). This can tell you when you are trading real movements and when you are just wasting your time trying to trade noise. John Ehlers has written some useful papers on this. Go to www.mesasoftware.com and download his papers there. They are quite useful and written for the non-technical reader so are very easy to understand.

CB
  • #190
  • Edited 8:16pm Dec 3, 2008 4:44pm | Edited 8:16pm
.
  • #191
  • Dec 3, 2008 4:54pm Dec 3, 2008 4:54pm
what do you think of this one?
Attached File(s)
File Type: ex4 JMA moving average.ex4   10 KB | 943 downloads
Trade2Win
  • #192
  • Dec 3, 2008 5:13pm Dec 3, 2008 5:13pm
Quoting supersub
what do you think of this one?
Can't open it. What is it?
  • #193
  • Dec 4, 2008 4:13am Dec 4, 2008 4:13am
To be honest I cant' see the difference!
Attached File(s)
File Type: mq4 JMA.mq4   11 KB | 1,145 downloads
  • #194
  • Jan 12, 2009 8:56pm Jan 12, 2009 8:56pm
found this thread via google. I'll say this is about the first really intelligent
discussion of trading I've found on the internet.

Anyways, codebreaker:
I'd like to hear a bit about your "clutter filter" or SnR as it would be.
I've been thinking and experimenting on the subject for ages and can't come up
with a decent idea. Mandelbrot has some great ideas in his book, but can only
be applied well after the fact.

At first I thought a simple kurtosis indicator would work - but again
that assumes a static price change distribution. So does everything else
I looked at. All great, if your distribution is a static bellcurve. The
market is anything but.

Are you taking your I[0] and doing a straight SnR calculation with your ullma?

If that's the case your clutter filter is relying on the fact that your ullma
doesn't follow congestion, no? So your clutter filter interacts with your smoothing
algorithms properties.
  • #195
  • Jan 13, 2009 5:26pm Jan 13, 2009 5:26pm
Quoting flyboy
found this thread via google. I'll say this is about the first really intelligent
discussion of trading I've found on the internet.
Thank you for the kind comments :-))


Quoting flyboy
Anyways, codebreaker:...
Yes and yes. I use the ULLMA to get a smooth price. I then do a spectral analysis based on the Goertzel algorithm and assign dominant cycles to "signal" and smaller amplitude cycles to "noise". Hence we get SNR. This helps to keep the system out of fake trades and whipsaws.

CB
  • #196
  • Jan 14, 2009 3:53am Jan 14, 2009 3:53am
Hi Codebreaker,
Did you do some work about CSSA ?
  • #197
  • Jan 14, 2009 1:38pm Jan 14, 2009 1:38pm
Quoting asm8086
Hi Codebreaker,
Did you do some work about CSSA ?
Hi asm

Yes I did. I programmed SSA for both Matlab and also in C for my trading platform. Fabulous technique but, like the Hodrick-Prescott filter we have also talked about, the method is non-causal. This means that the right-hand endpoints keep "changing their mind" as new data arrives. This means that the indicators constantly re-paint and thus they are useless for trading live.

I have scoured the forums for info on Noxa's "Causal" SSA but found nothing. I can't even find feedback from anyone who has bought their system and tried it. :-((

CB
  • #198
  • Edited 4:17pm Jan 14, 2009 4:14pm | Edited 4:17pm
Hi CB,

I found the Post #170 from Noxa in response to my question #169 very interesting and helpful :

Quoting Noxa
Have you ever thought that a causal version of CSSA could be learned from SSA with a recurrent neural network such as an Echo State Network?

Noxa
Like you I have a matlab version of SSA and I coded it in C# for my trading plateform but I don't know for the moment what to do with the SSA coefficients in an Echo State Network... (I have a lot of ESN codes, but I didn't take the time to understand it correctly)

Have you incorporated this network in union with SSA ?
  • #199
  • Jan 15, 2009 3:45am Jan 15, 2009 3:45am
Quoting asm8086
Have you incorporated this network in union with SSA ?
Of course, I saw this enigmatic answer. I have neither the time nor the inclination to expend any effort on this. One could approach this a hundred different ways. If I get an interested MSc student I might get them to take a look. But I don't think that the method will be that useful.

IF, and I mean IF, you hold the view that the market can be broken down into sinusoidal cycles which may persist for a short time into the future, then a spectral analysis followed by cyclic extension (with the appropriate phase) will give you all you want to know.

I cannot see how SSA can be "causal". SSA is a spectral analysis technique. The only way such techniques can be predictive is by cyclic extension.

CB
  • #200
  • Jan 15, 2009 4:27am Jan 15, 2009 4:27am
Hi codebreaker,

I would like to ask you whether you know about the invariant concept that is applied in modern trading? It is based on stochastic process and multiple time frame analysis. Invariant concept comes from mathematics that something does not change undet a set of transformation. Can you share your knowledge on this?
THANK YOU.
Trade2Win
1 trader viewing now