Magic trade strategy

There is an interesting dicussion on the forexfactory forum, of a method to predict the current leg of the Zigzag indicator.

http://www.forexfactory.com/showthread.php?p=7095082#post7095082

They have incredible results. Maybe that could be a new trade strategy for Zorro?

Interesting.

I'm sure it's possible and would love to give it a go but haven't the coding or mathematical knowledge at this point in time unfortunately.

I was also notified by some people about that discussion of using fractal sequences for predicting ZZ turning points. I have even read some part of the thread, but I'm not sure what to think about it.

Pro:

- the author, gg53, got miraculous results with his system
- he seems to be a honest and serious person

Contra:

- no backtest - the miraculous results are just anecdotical
- apparently no one else could reproduce them so far
- the algorithm doesn't really make sense

I think a system that works for its author, but for no one else, is not really useful. If someone wants to try it nevertheless, I'll help. But even for an experienced programmer, scripting this algorithm would take at least a week.

And I haven't understood his first post where he starts with his sequences LOL

'short 4-string linear sub-sequences end on X'

Does he anywhere explain how he created this sub-sequences?


PS: Ah. I've found it:

If X(n) is a sequence, a linear subsequence is any subsequence of the form X(an+b) where a>0, b>= 0

FYI. Here is a more sensible offshoot from the original thread;

http://www.forexfactory.com/showthread.php?t=464948

Looks like this theory is gaining traction and may very well suit our beloved Zorro?

Unfortunately, it'll take someone much smarter than me to do anything with it here

I haven't followed much of the thread, but as far as I've understood, even after 6 months hard work no one could really reproduce or backtest this method. Nor makes it much sense from a theoretical or mathematical point of view.

Still, if it can ever be confirmed that the method really works, it can certainly be coded with Zorro and made available as a Z system.

To me, there are many reasons why this logic will never ever work. But the most direct and noteworthy reason is that the product of price change that we can see as graphs is a result of human behavior. Now what is the similarity of human behavior and the phenomenon of fractals in nature? Nothing. Human behavior is irrational while fractals are. Fractals can be mathematically justified and mapped out. Human behavior cannot be mathematically mapped out. It would take an infinitely complex system with infinity of variables. Trading algorithms are simple mathematical equations with constants.

In fact, the very mistake of undertaking this whole project of "prediction" of price movement - a "Holy Grail" is irrational human behavior.

That said,the only area that is worthy of research in trading is exactly the opposite. It is to study what shouldn't be there - the inefficiency of the markets and develop trading approaches accordingly. So instead of looking at fractals and asking what's next, we should be asking what is out of place? What should not be in this fractal?
...this just gave me an idea...

Related to complex models and fin. algorithms and the issue of simplification:
youtube.com/watch?v=ed2FWNWwE3I#t=12m04s

Thanks for the video Finstratech, I enjoyed it! I see there are many others related as well.

Hello !
I've Just joind this forum.
It's me you're talking about...
You can find the system details here (we moved from the original thread):
http://www.forexfactory.com/showthread.php?t=464948
...and it's not working just for me.

G.

Thanks for joining. Is there already some backtest result? I see a trade journal in the new thread, with 40% loss after 150 trades - is this from live trading the method?

As far as I understood your method, it's more an intuitive system and not based on a mathematical model or theory. So aside from live trading, a backtest would be the only way to see if it's useful.

If you have problems to backtest your system due to some MT4 limitations, maybe we can help.

Post #1 (in that forum) topic "Double within a month" explains the history of the results and the bugs found in that EA/test-run in the TradeExplorer.
After bug fixing we managed to double the account within a month - as we targeted.
I don't believe in backtests for performance - only for bugs finding and system behaviour analysis. Forward testing is the ONLY way to see real results and performance of a system.
There is nothing intuitive in that system - but who am I to judge your understanding...

G.

Thanks for the explanation. You're right that it is not trivial to correctly test a system - but not testing it at all is no solution either. I don't think that "double the account within a month" is a serious test. Anyone can double an account in a month, with any system, just by using a risky stop/profit ratio and a bit of luck. Zorro has a tutorial about those methods.

Have you ever seriously live traded your system? If so, with which result?

My offer still stands to provide you with the software tools for a real forward or back test of your system.

We are trading this system live for a month now (which is not enough).
The results are quite good (82% wins, 4.7% DD) - but that's not indicative because of that short time.

The ForexGT_Arrow system is just a small part of the strategy: We use CurrencyStrength & Correlation to decide which pair to trade, Volume & Volatility for "IF" to trade, and ForexGT_Arrow for entry.

Many thanks for your offer to help - We own a software company - so I don't think that we currently in need of back/forward testing assistance.

G.

Hi, is there some public record of your live trading the magic strategy, or of anyone else trading it? Or is the trading also top secret?

If you can only test your system with live trading, why do you add the "CurrencyStrength & Correlation" and the other stuff? It seems to me that for finding out if your method works, you should trade the pure method instead of adding lots of other stuff that affects the result.

Petra, my take on those comments:

Originally Posted By: gg53
The ForexGT_Arrow system is just a small part of the strategy: We use CurrencyStrength & Correlation to decide which pair to trade, Volume & Volatility for "IF" to trade, and ForexGT_Arrow for entry.

Is that Currency Strength and Correlation and Volume and Volatility form part of the entry criteria, hence it is part of the strategy, and not about measuring the results.

No, they are not part of it, gg53 has explained his method in very long threads on the FF forum.

I only wonder why it is so complicated to do a simple test. The method just predicts ZZ points. So you could easily test it by going long at a low point and short at a high point. If the method has any merit, the results should be positive. I wonder why after almost a year, still no one can show positive test results.

Well this is a valid point. It's no secret that professional traders do not use indicators, and for good reason, almost none of them have an edge. So why worry about it?

Hmmm, I dont think that this system has ever worked.

The public trading on the FF forum had bad results, except for one short period in the middle with lots of small wins. But those wins are obviously faked with a high risk reward ratio. You can see this because the profits have always the same size. If they were really from the system, they would be all different because ZZ legs have very different lengths.

All this looks very fishy. There are some poor guys on the FF forum who still believe in that system and still try programming it for almost a year now...