Digital ACSTrend - page 27

New comment
 

questions

Well, if it works do not fix it.

First look at the original thread and what was done there. The code is very interesting. I doubt that my mod is successful.

Anyway everyone has his own choices. I do not think that there is an universal solution. My personal choice is Rapidminer, as I am incompetent in programming and it is open source and a lot of institutional players use it. See a nice video how to build a classification model.

http://www.neuralmarkettrends.com/2010/09/30/rapidminer-5-0-video-tutorial-11-%E2%80%93-pattern-recognition-landmarking-plugin/

if you want to test it use the previous version 5.0 and not 5.1

This topic exceeds the scope of this thread. In FF in the thread of optimized trend trading I suggested to develop together a model. But nobody is really interested.

Anyway I believe that it is possible to build a good classification model that can be used together with any mt4 strategy.

Real professionals do not use MT4:Powerful PC, Linux, Custom Linux kernels, C or C++ code, custom API with direct market access.

We are far away from this world. Anyway in the next years we may see an accessible High Frequency infrastructure provided by third parties.

If we see that we may then ask the question is there still an edge into all this LOL?

The fun point is that all this may perform sometimes as well (as bad) as an amateur MT4 strategy. We are going to hear stories for loosing money at high frequency LOL. This story will be a nice one.

 

Some pips

Here I want to give an example that a good time frame to trade is where we have low fractal dimension. This is giving us the edge that our directional algorithms are going to work.

There are many systems around but this is fundamental how to use the market memory. Today many of my sophisticated neural nets failed, but my directional algorithms performed fairly well.

Files:
sihh.gif  34 kb
 
John Last:
Here I want to give an example that a good time frame to trade is where we have low fractal dimension. This is giving us the edge that our directional algorithms are going to work. There are many systems around but this is fundamental how to use the market memory. Today many of my sophisticated neural nets failed, but my directional algorithms performed fairly well.

Hi John

I am new to fractal dimension and I was looking up the FDI indicator. Are they similar?

 

Hello guys

I do the metatrader update to version 392 and i have the same problem !

There is surely a problem with ssa indicators. When new ticks comes it re-calculates every time and it consumes 100 % of my processors !

With our old metatrader everything works, so :

DO NOT MAKE THE UPDATE !

Maybe a coder master could take a look This will be great

 
Big Joe:
Hello guys

I do the metatrader update to version 392 and i have the same problem !

There is surely a problem with ssa indicators. When new ticks comes it re-calculates every time and it consumes 100 % of my processors !

With our old metatrader everything works, so :

DO NOT MAKE THE UPDATE !

Maybe a coder master could take a look This will be great

Well, doesn't really matter which version, MT cannot handle complex indicators. Probably it is better to move one to some decent platforms like NinjaTrader etc....

 

Big Joe

For statements like that (that something is wrong with the "new" metatrader 4 versions) and statements like this one (this post of mine) some people got banned from mql forum

As I remember the error you are describing was also happening in 387, then they corrected it, then they un-corrected it, then they will correct it again (eventually) ... and so on and so on ...

Sadly, it is the same manner as in metatarder 5 : you never know when will some past error resurface again with some "improvement" of metatrader and then you have to wait a week or two to be able to do anything meaningful

For now all we can do is avoid updates and wait till the "gods" of metarader decide to release only versions that are tested and not the ones that will show that something is going on

___________________________________

PS: isn't it interesting that the changes and errors they are making in metatrader 4 are almost the same as they make in metatrader 5? So will it be an up-grade or down-grade?

 

Lyapunov exponent

Hi all,

Today we had a very intensive day LOL. Jaguar1637 developped a trial for calculation of the Lyapunov exponent (he prepared some more sweeties I hope he will release to the general public soon). Well still we are not sure for the implementation. However still we have something.

Here it is what he published in the thread of optimized trend trading.

Forex Factory - View Single Post - Optimized Trend Trading

I made a little modification by using averaging with MA.

More information about the lamda (Lyapunov exponent) you can find here:

4.3 Lyapunov Exponent

Well the interpretation from technical point perspective is quite simple. When it goes below zero prepare for an impulse.

I add a screen shot of what we have.

Files:
lyapunov.gif  24 kb
lyapunov_jytv3.mq4  4 kb
 

...

John Last...I am very good at detecting patterns in a price series...I look at the picture and do not see any single useful pattern....What do you see over there, that is making you excited?

 

Chaos indicators

I think you should read this:

Chaos and Stock Market

It can be simple but It cannot be any simpler.

I add a shot. I consider this pretty obvious. There are not any directionnal patterns. But those indicators are not directionnal, as the ATR e.g.

Anyway do not forget this, it is just a prototype without any guarantee for consistence. However in another forum I am really in war with price actionists "and technical pattern specialitsts", some of them are real professionals. They just do not get it. Anyway that is not my problem. I am not looking for any kind of consensus. Consensus sucks because it is useless in market environment!

However it is worth to know that there is a major contrarian opinion that those patterns are worthless.

However I prefer to count on my system. Is it a holy grail?

Of course not, just the market was predictable at that time.

Files:
patterns_1.gif  25 kb
patterns2.gif  29 kb
1...202122232425262728293031323334...43
New comment