Magic of ADX

uasish · #71 28th June 2007, 08:31 PM

Quote:

Originally Posted by CreditViolet

Nice!

A truer test of the system would be on intraday data, try it out

Thks,prepearing for the same as suggested.These are only different parts,until i get the whloe you have to be lonely for a while.

CreditViolet · #72 28th June 2007, 08:39 PM

Quote:

Originally Posted by rajak1981

Agreed: they have been around for a while.
as mentioned in my post Neuroshell does both Ga+NN.
There is also another free program called wealthlab which does NNs. I have atleast 2-3 other programs whose trial versions have now expired. Most of them cost too much hence I haven't bought them.

Your best option for NN would be NS or Matlab, for training NN interpreted languages would be too slow.DLLs would be more efficient.

Quote:

Originally Posted by rajak1981

Whats the cheapest ga+nn solution available?
Please post a price + link. Should help me decide whether my project is financially viable. I'm a bit of a theoretical person and I often start up projects without working out concrete details. Refer to my old post I was crazy enough to verify cycle length/ob/os levels for rsi.

Take a look at Alyuda Products, this very much depends on the data you will be training on.I use intraday data and TS DLL interface is good enough for me currently

Quote:

Originally Posted by rajak1981

Are you a coder? Have you coded your own self adaptive systems? If so I would love to get a headstart on coding my own ga/nn by dicussing with you.

[i]http://www.traderji.com/attachment.p...1&d=1183035268

Umm ya, I am a coder.I havent coded my own self adaptive functions yet, I use commercial ones.

Quote:

Originally Posted by rajak1981

Have you tested t3 fibs ? Does it come with a backtest?
I Haven't tested T3 fibs in detail so I do not know how well it works: I only reversed t3 to prove that I could reverse it.
Do you own t3 fibs/tradestation? What data feed?
Does t3 fibs come with a backtestable set of rules? It seems like a lot of hocus pocus on their side. I don't even have t3 fibs. Still trying to find a copy. Not crazy enough to pay 5000-10,000$ for a system.

I hve owned the t-3 system since mid 2004, its not exactly a documented system to backtest, you will have to check the accuracy by keeping records manually.
Data feed is eSignal.

Quote:

Originally Posted by rajak1981

What is the attatched nn picture? Your own nn?
How much does it cost? Link please and performance report.

Yes Sir.It doesnt cost anything bcos its not for sale yet! I think I posted a perf report further down below.Chk it out

Quote:

Originally Posted by rajak1981

Agreed:
Everything has been around for a while.
How much do the systems on the clayburg site cost?
I will certainly have a look at them. I'm applying for the free trial as we speak. Thanks for the link.
Check out this link for Indian markets , has been around for atleast a few years I guess. Don't know dates.
http://www.stratstar.com/

I'll be content to market something low end along the lines of statstar which just gives live signals or charts without having to give away my system to every buyer.

I am aware of stratstar, who knows what they do but they dont post any detailed statistics of their systems.

Quote:

Originally Posted by rajak1981

The multifib + pattern recognition systems mentioned in the most recent post are not finalized yet because my key ingredient is missing. An adaptive ga+nn. If you can suggest one at a low enough price point then I will look into it. Something that can integrate with AB would be nice. Perhaps what you are showing in your picture.
Testing each pattern by hand using simple rules would be a waste.
My theory is: it might just be better to build the ga+nn and let it rip through old sample data to help me find rules/ideas for trading that work.

Well I actually tested by hand before moving to the advanced stage and it has been very useful

Quote:

Originally Posted by rajak1981

For t3 fibs: I still need to work out the confluence collapse margins + exact t3pivot rules. I am using a customizable set of pivot rules right now. I will need to integrate the fib confluence pain with the other panes.
So right now my system is not a perfect replica nor is it fully integrated
Will post pictures later.
My confluence system is just a brute force chart which plots a crazy amount of levels! no confluence detection/resolution finalized yet. I have studied the parameters that T3fibs provides, still need to convert them into code.
How do you figure out which fib level to keep and which one to drop? Have you worked out a full confluence system? Do you have code? Interested in sharing? for a price? How good is it? Please post a picture of filtered levels+ trading system+ backtests.

It took me 1.5 years to fully replicate and better the T-3 levels, coding it is indeed a lot of work. Sorry, too much work, dont want to give it away, it pays for itself quite well.
Pics n Results later.

Quote:

Originally Posted by rajak1981

I hope to correspond with you in the future. I have a whole bunch of questions which I believe you are well equipped to answer.

Here is one question for starters

. You might be able to answer it since you actually have a copy of t3 fibs.

Raja (6/25/2007 7:07:41 PM): Raja: http://www.nss-t3.com/newvideo/pivot..._expected.html this is not a bill williams fractal and not a gann swing pivot or a one bar reversal what is it exactly? can anyone make out?
Raja (6/25/2007 7:07:49 PM): figure out the mathematical rule that causes the pivot to plot

Cheers,
Raja

I dont use the Pivot Paintbar, well I have my own thing for tht as well.It calculates Pivots more on the basis of Pivot Strength which is also behind the confluence indicator.

Rgds

oxusmorouz · #73 28th June 2007, 08:50 PM

Quote:

Originally Posted by rajak1981

I have evolved my own versions/improvements/mutations to try and fit emh with technicals.
Perhaps I wasn't precise enough
Take emh/markowitz/capm add it all up

Oh really?

With due respects, what am I to call it but a hoax? You fail even provide an example

Quote:

Originally Posted by rajak1981

do I really need to read emh or do you really need to read my posts?

That is for you to find out, I guess. I should think the contrary

Quote:

Originally Posted by rajak1981

Proof by contradiction.
Adx followers are the ones under pressure.
Refer to my first post.. Go through all criteria required to prove adx is magical.
If you wish to win this debate.

I think I have replied to this earlier. Have you gone through post no.101?
I guess it is time for my post to get iggyd ..sniff

Quote:

Originally Posted by rajak1981

Take two good ideas. parents, cross over, mutate.
come up with a new idea
assume technicals work partially (i.e price action repeats/patterns etc)
assume emh works partially (i.e market is usually efficiently priced)
find the middle ground.
What is the middle ground???
I guess I would get more respect if I went and got an mba in finance or something.
Thanks for the pointer.

If you are still confused
I am the child of my mother and father.
I am not my mother.
I am not my father.
My mother and father are very different human beings.
I do not look/behave exactly like my mother, nor do I look/behave exactly like my father.
In some ways I am even different from both of them.
I still give credit to both my mother and father for who I am and I respect both of them deeply and equally.

Nice story but where does it lead? Looks to be more delusionary than explanatory. Please answer my question.
And yes, just to make 1 point clear, I'm still in my teens and not even a graduate yet

oxusmorouz · #74 28th June 2007, 09:58 PM

Quote:

Originally Posted by rajak1981

What is your point?
You want to say adx works?
BAV already did that?

You can just say adx is not magical and the debate ends

Asish/BAV and I all agree on this simple point.
The only conclusions thus far have been neither adx nor anything else is magical.
If you want to end there. Fine, then you have made no point.

The responsibility to prove adx's magic lies with the DEFENDERS of this thread.
They have to post code. i.e the magic spell.
I'll only post code for garbage which is one notch above garbage posted by you. So do your best. Show some garbage which uses adx with code and I will improve it by substituting the adx filter with something else. Hence I will prove that adx is not magic. This is just my hypothesis. It is entirely possible that you will post a system which employs adx in such a cool fashion that nothing I do would be able to replace adx. In which case everyone including me learns something new.

Your limitation when you post code is that you must use adx, my limitation is I am not allowed to use adx. I am quite willing to risk defeat in a small thread to learn something new!!

The whole purpose is not so much about being right or wrong but about learning new ways to use adx. Do you wish to compete or back down now?
If you wish to compete post the full code.
Regards,
Raja

My my, you seem to have entirely missed the point.

When did I ever talk about ADX being magical or when was I a die hard fan of it in the 1st place? My question was attacking your theoretical mumbo jumbo (not based on evidence, example or anything of that sort) which tried mixing EMH with technical analysis and bases mechanical trading systems on the premise of CAPM, which seems to be rather, utterly ridiculous to me, lacking basic understanding of the concept itself. You even claimed to have made a "dynamic coding" of allowing the capital asset pricing model into your amibroker. Would you care to give your code out? Ofcourse you won't! It's your little secret

(like it exists! bah!)
As a wise man told me in one of his earlier posts, it is best to take a contrarian approach to a problem, taking that the entire thing is wrong as the null hypothesis. Care to refute this null hypothesis of mine with your code (just like you asked for mine, in your previous post) ?
Or do you simply wish to back out?

uasish · #75 28th June 2007, 10:05 PM

Raja,
Dev started this thread saying ADX works better & Traderji gave his input.
Your this post
http://www.traderji.com/95445-post8.html
was against that.
Your this post
http://www.traderji.com/95520-post27.html
even without backtesting you became sure ADX works.
http://www.traderji.com/95478-post20.html
After backtesting you were convinced,ADX works,so where does we stand & you stood & now stand.
I dislike cloudy things.

Asish

uasish · #76 28th June 2007, 10:12 PM

Raja,
I ,we ,all believed ADX worked ,oxy suggested us to prove it by test results.
Eventually he did that,you are changing your stand,all these are not only disliked but extremly undesireable .

Asish

uasish · #77 28th June 2007, 10:35 PM

Raja,
Oxy is Ajay,my teacher,9 yrs youger than my son,you are yet to match him.

Asish

oxusmorouz · #78 28th June 2007, 10:38 PM

Quote:

Originally Posted by rajak1981

P.S I was only applying occam's razor to adx
" This is often paraphrased as "All things being equal, the simplest solution tends to be the best one." In other words, when multiple competing theories are equal in other respects, the principle recommends selecting the theory that introduces the fewest assumptions and postulates the fewest hypothetical entities. It is in this sense that Occam's razor is usually understood."

Yes yes, but does not assist you in anyway. The focus of my arguement was not to support the ADX but to question your premise. I have made this very clear.

Quote:

Originally Posted by rajak1981

Do you wish to prove anything regarding adx being magical??

The answer to that, I'm sure, is already given. Perhaps we should read other's post before asking them to read ours?

Anyway, what I wanted to learn from this thread is achieved. Good night.

SavantGarde · #79 28th June 2007, 10:39 PM

Just As So Many Assumptions & Variables People Are Trying To Put Together To Come Up With The So Called "Holy Grail"

At The End Of The Day It Is The Human Mind At Work Or Play.

Just As "Automatic" Doesn't Mean If Something Is Automatic, It Will Repair Itself.

Here Is An Example Of Actual Trades Of A Logical, Aware, Awake & Extremely Focussed Human Mind.....Please Don't Waste Time, Trying To Look For A System In It.... Cause There Is None, But That Doesn't Mean One Cannot Derive Trading Strategy From The Actual Trades Posted, Maybe It Will Help With Scaling Up or Scaling Down & Not Hesitating In Converting Unfavourable To A Favourable.

From Somebody Who Entered The Market On 17th May 2004, Does It Ring A Bell, Let Him Know If It Doesn't...

Yessss.............Over A Period Has Devised A System..........Though It Is Constantly Being Evolved & The Results Of The System Is Posted Elsewhere In This Forum.

What Is This Person's Qualification: Dropped Out From 12th, Taught Aspiring Doctorates, Is An Architectural Consultant (& Couldn't Practice Being An Architect Cause Didn't Have The Finances) But Lack Of It Doesn't Stop Chartered Architects From Seeking Him Out....Therefore The 'Thappa' Of Architectural Consultant & Now A Full Time Trader.....Cause Chasing Dues, Seemed To Be A Full Time Job, Though Architectural Work Is Still Practised But Only If It Meets His Terms & Conditions....Which Is Full Fee In Advance..

Has Immense Respect For All Those Warren Buffets of India Who Applied for Infosys IPO & Have Held On To It.....

Mission Is To Convince People, Derivatives Is Where The Real Money Is & Probably The Safest Instrument To Play The Markets, Hears Murmurrrr....Of Disagreement......Well, Well, Let's Hear The Nature Of Disagreements.....& Maybe Then, Shall Present Supporting Views... In An Appropriate Thread....Do Not Want To Populate This Thread With "Why Derivatives Is The Safest Instrument"

SavantGarde

Niranjanam · #80 28th June 2007, 10:45 PM

Cognitive Dissonance ?

Similar Threads for: Magic of ADX
Thread	Thread Starter	Forum	Replies	Last Post
The ADX(Average Directional movement index)	sh50	Technical Analysis	15	17th January 2007 11:54 AM