magic shift | Page 5 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Magic numbers question 12 replies

What does GMT shift and GMT shift Sun do in this indicator? 0 replies

magic is back! 11 replies

Magic Numbers for Mutiple EA 9 replies

Attachments: magic shift
Exit Attachments

magic shift

  • #81
  • Dec 1, 2009 9:15am Dec 1, 2009 9:15am
Did i already mention that this strategy is not only dangerous but also extremely boring?
  • #82
  • Dec 2, 2009 12:07pm Dec 2, 2009 12:07pm
Quoting smjones
Here is an idea I have been using on this in testing manually that seems to make it safer without cutting profit, in fact enhancing the profit.

The attachment is a series of slides with notes about what I am doing with an unusual trend line. Perhaps you could incorporate something like this into your EA. So it will trade the signal from the AC indicator if the Trend line is ok. The point here is this is not a normal trend line, because the trade is self is adjusted bar by bar. Take a look and see what you think. I have not really researched this...
Scott, thanks for taking the time to post this. I have found many of your posts to be real 'gems'.

Dan
  • #83
  • Dec 4, 2009 9:33am Dec 4, 2009 9:33am
Quoting andante9
Scott, thanks for taking the time to post this. I have found many of your posts to be real 'gems'.

Dan
Just in case you guys were wondering if this could be profitable. This is just trading the unconventional Trend lines alone this week and not even using the AC method. I believe it would make an excellent ratchet.
Attached Image (click to enlarge)
Click to Enlarge Name: ScreenHunter_032.jpg Size: 84 KB
  • #84
  • Dec 4, 2009 11:06am Dec 4, 2009 11:06am
Quoting 7bit
I trade it forward only since a few weeks..., it just destroyed 3 weeks gains (~300 pips) in a few hours, just before the weekend ... This system is dangerous and i'm still trying to find ways to make it safer without keeping it from making profits!
Would you recommend trading it say from Tues London open to Friday London close? I know it may change testing results, but just thinking of potential profits, without the squeeze of impending weekend if presently in a negative position. I've been following this thread and I have to say ... Geeez! People sometimes complain that not enough 'lurkers' contribute, but what can a regular guy say to a crowd of Einstein's LOL
  • #85
  • Dec 4, 2009 12:26pm Dec 4, 2009 12:26pm
Quoting ttradr
Would you recommend trading it say from Tues London open to Friday London close? I know it may change testing results, but just thinking of potential profits, without the squeeze of impending weekend if presently in a negative position.
The funny thing is, *sometimes* it is profiting just from these crazy and seemingly unpredictable movements. Today it went short for just 2 candles, catching half of the nfp move:
Attached Image (click to enlarge)
Click to Enlarge Name: magic_h4.gif Size: 20 KB


while the H1 version didn't have this luck:
Attached Image (click to enlarge)
Click to Enlarge Name: magic_forward.gif Size: 17 KB
  • #86
  • Dec 4, 2009 1:15pm Dec 4, 2009 1:15pm
Trying 2 different concepts on 2 different brokers. Second one I had changed from -61 to -66 yesterday.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: magic_shift diff 01.gif Size: 17 KB
Click to Enlarge Name: magic_shift diff 02.gif Size: 20 KB
  • #87
  • Dec 4, 2009 3:35pm Dec 4, 2009 3:35pm
Damn man bad timing huh? It seems like this system is just a coin toss to me
  • #88
  • Dec 4, 2009 4:23pm Dec 4, 2009 4:23pm
The concept is brilliant. Optimization is key to it's profitability or otherwise. As 7 bit has stated, this is certainly not a system for newbies.
Thanks for sharing 7bit.
  • #89
  • Dec 5, 2009 5:11am Dec 5, 2009 5:11am
7bit,

why exactly did you chose AC as the basis for correlation search?
AC is kind of a smoothed 2nd difference that uses two SMAs for smoothing, for some reason with 5 and 34 period lengths (probably by empirical optimization). Considering this, your system is actually a three parameter system - two filter lengths and a lag. So far you keep these filter lengths fixed and vary only the lag. Have you experimented with optimizing for these two parameters, too? My guess is that varying the filter lengths has, to some extent, an influence on the significance of correlation peaks.
  • #90
  • Dec 5, 2009 9:07am Dec 5, 2009 9:07am
Quoting smjones
Just in case you guys were wondering if this could be profitable. This is just trading the unconventional Trend lines alone this week and not even using the AC method. I believe it would make an excellent ratchet.
Question to smjones:

If you look at the attachment where the green arrow points, I would say that the close would be on the reversal side of the trendline. The red arrow points to the previous low of the bar where the trendline would have went through. How come, as it seems, you didn't close or reverse the trade?
This is just one example. I see plenty more, but I suppose there is more to this strategy of adjusting the trendline.
Thanks.
Attached Image
  • #91
  • Dec 5, 2009 9:51am Dec 5, 2009 9:51am
Quoting fewhills
Question to smjones:

If you look at the attachment where the green arrow points, I would say that the close would be on the reversal side of the trendline. The red arrow points to the previous low of the bar where the trendline would have went through. How come, as it seems, you didn't close or reverse the trade?
This is just one example. I see plenty more, but I suppose there is more to this strategy of adjusting the trendline.
Thanks.
I was wondering if anyone would see this. LOL Ok here are the rest of the rules.

1. If the bar that you would move the trend line to is an inside bar, keep the same level for the trend line as the anchor bar.

2. if you have a clear Pin Bar, that becomes the exit close and the new trend line starts where it should have.

3. if you have a stopping volume low or a stopping volume high, that becomes the exit close and the new trend line starts where it should have.

That is all I want to say on this thread about this, because I feel like I am hijacking this thread. Also, This was not designed to be a complete system, just some musings in my head. It certainly could develop if some one wanted to as a stand alone method.
  • #92
  • Dec 5, 2009 1:43pm Dec 5, 2009 1:43pm
Quoting jtmoore007
Damn man bad timing huh? It seems like this system is just a coin toss to me
Yes, it is like a coin toss, but some of the coins are biased. Every other trading system is also nothing more than tossing more or less biased coins.

The H4 system has made 6 trades since this thread started. It is absolute too early to draw *any* conclusions from only 6 trades. Maybe after 60 trades we can discuss whether the coin we selected was the wrong one, but not after only 6 trades.

Additionally we may just witness the beginning of a major trend reversal and maybe the "rythm" of EURUSD also just changed. Time will tell.
  • #93
  • Dec 5, 2009 2:03pm Dec 5, 2009 2:03pm
Quoting Mechthildche
7bit,

why exactly did you chose AC as the basis for correlation search?
AC is kind of a smoothed 2nd difference that uses two SMAs...
There is no special reason. I assume that the internal SMA periods have no meaning at all (at least not for what i'm doing here), other than controlling the smoothness. I have something similar and a few months older here: http://sites.google.com/site/prof7bit/autocorr
It uses the following code instead of AC:

Inserted Code
double getHistoricDirection(int shift, int lag){
   if (lag == 0){
      return(0);
   }
   
   int sign = 1;
   if (lag < 0){
      lag = -lag;
      sign = -1;
   }
   double o = iOpen(NULL, 0, shift + lag + 2);
   double c = iClose(NULL, 0, shift + lag - 2);
   return (sign * (c-o));
}

It will just use the price change during 4 bars, so it basically uses the first difference. The smaller the number of bars, the more often it trades, the effect will stay the same, only the profitability decreases. In the current incarnation of the magic_shift EA i chose AC simply because it was the first thing that came to my mind when i needed something that would somehow "go up and down with price" and wouldn't be too fast (not too many trades). I am absolutely sure i could also use Stoch or RSI or OsMA or almost any other oscillator, be it first or second difference, it wouldn't matter.

Maybe i could replace it with something with adjustable "smoothness" to control (and optimize) the trade frequency.
  • #94
  • Dec 5, 2009 2:45pm Dec 5, 2009 2:45pm
Just asking, what will you do when it starts to fail ? You found out it works with 61 because you have backtested it. But as soon as it becomes unprofitable it will be enough to lose some of your profits and, moreover, not to be able to find a new configuration becuase you don't have enough data for backtesting.
  • #95
  • Edited 3:43pm Dec 5, 2009 3:32pm | Edited 3:43pm
Quoting Pollux
Just asking, what will you do when it starts to fail ? You found out it works with 61 because you have backtested it. But as soon as it becomes unprofitable it will be enough to lose some of your profits and, moreover, not to be able to find a new configuration becuase you don't have enough data for backtesting.
This is a difficult question. I have no answer. Maybe turn it off and wait a few weeks, see how the old shift would have behaved, try to find a new one. Maybe this trading system is only of academic use, to demonstrate some interesting market behaviour and cannot be used profitably at all or maybe there is always only a small time window for trading it profitable.
  • #96
  • Dec 5, 2009 5:11pm Dec 5, 2009 5:11pm
fuck ODL
  • #97
  • Dec 10, 2009 6:44am Dec 10, 2009 6:44am
please , have you any indicator for "point's 61" on AC indicator.Thank's and sorry for my bad english...
  • #98
  • Dec 10, 2009 7:21am Dec 10, 2009 7:21am
Quoting paul czech
please , have you any indicator for "point's 61" on AC indicator.Thank's and sorry for my bad english...
I'm not sure if i understand the question correctly, but if you use the EA from Post #1 in this thread it will draw a vertical line exactly at that position. If you then additionaly attach the indicator to the chart you will see it.

PS: the shift 61 on h4 seems to have stopped working at the exact day i posted this thread here and the 192 on H1 is more or less break even. I'm still running them on the demo and record their profit curves but have stopped mirroring the trades on my live account.
  • #99
  • Dec 10, 2009 7:54am Dec 10, 2009 7:54am
Quoting 7bit
I'm not sure if i understand the question correctly, but if you use the EA from Post #1 in this thread it will draw a vertical line exactly at that position. If you then additionaly attach the indicator to the chart you will see it.

PS: the shift 61 on h4 seems to have stopped working at the exact day i posted this thread here and the 192 on H1 is more or less break even. I'm still running them on the demo and record their profit curves but have stopped mirroring the trades on my live account.
yes,it's ok , my broker have only mini - lot .....setup in EA was on micro lot... and EA now it's ok
  • #100
  • Dec 10, 2009 8:24am Dec 10, 2009 8:24am
Quoting paul czech
yes,it's ok , my broker have only mini - lot .....setup in EA was on micro lot... and EA now it's ok
If it doesn't start working again with 61 in a few weeks i'll start trying to find a new shift. This performance drop more or less coincides with this Dubai thing, the last NFP and the possible EU trend reversal we are seeing. Maybe some things have fundamentally changed.
0 traders viewing now