THE IDEA EXCHANGE - page 31

New comment
 

There was a long discussion on the trend-flat filter in the thread, I'll add my 5 kopecks to this topic.

I use the comparison of ATR indicator values for two periods (fast and slow) to distinguish a trend-flat. The logic is as follows: slow period most probably includes trend and flat periods, respectively, if fast ATR is not less than slow one, it means that most probably a trend period has begun. On M15 I use period value 4 for fast period and 96 for slow one. I.e. I compare the ATR of the last hour with the ATR of the previous 24 hours. Certainly, if one optimizes on separate segments then period's values can be more effective, but in my opinion my parameters are more or less universal.

   double _atr=iATR(NULL,0,4,1);
   double _atr1=iATR(NULL,0,96,1);

if(_atr>=_atr1...); // тренд
 
The problem is how to find the best averaging period. If we assume that the market picture (I don't mean trend or flat) will continue rather than change, then we have a chance to build a profitable Expert Advisor. The idea of adapting the wrecker is not new, though I have not succeeded as you have ( I don't have to answer and spit out). I want to try to automatically change the dummy period to keep the price between two and three sigmas, if someone has tried it, please give me a hint.
 
Sancho77:

There was a long discussion on the trend-flat filter in the thread, I'll add my 5 kopecks to this topic.

I use the comparison of ATR indicator values for two periods (fast and slow) to distinguish a trend-flat. The logic is as follows: slow period most probably includes trend and flat periods, respectively, if fast ATR is not less than slow one, it means that most probably a trend period has begun. On M15 I use period value 4 for fast period and 96 for slow one. I.e. I compare the ATR of the last hour with the ATR of the previous 24 hours. Certainly, if one optimizes on separate segments then period's values can be more effective, but in my opinion my parameters are more or less universal.

The same SuperTrend
 
ivandurak:
The problem is how to find the best possible averaging period. If we assume that the market picture (I don't mean trend or flat) will continue rather than change, then we have a chance to build a profitable Expert Advisor. The idea of adapting the wrecker is not new, though I have not succeeded as you have ( I don't have to answer and spit out). I want to try to automatically change the dummy period to keep the price between two and three sigmas, if someone has tried it, please give me a hint.
And what indicator will the change of the MA period depend on?
 
Idea number 1.

It's more in the spirit of those researching Martingale.

Create an Expert Advisor for a tester and see what happens.

Trading Criteria:

0 min - Buy = 0.1
2 min - Buy = 0.2
4 min. - Buy = 0.4
8 min.- Buy = 0.8
16 min. - Bai = 1.6
32 min. - Bai = 3.2
64 min. - Bai = 6.4
128 min. - Bye = 12.8
256 min. Bye = 25.6
 
alex12:
Idea number one.

Rather close in spirit to the Roulettes.

Create an EA for a tester and see what happens.

Trading Criteria:

0 min - Buy = 0.1
2 min - Buy = 0.2
4 min. - 4 min. - Bye = 0.4
8 min - Bye = 0.8
16 min. - Bai = 1.6
32 min. - Bai = 3.2
64 min. - Bai = 6.4
128 min. - Bye = 12.8
256 min. Bye = 25.6

512 min = Uncle Kolya
 
Vinin:

512 minutes. Uncle Kolya.


:))

alex12, what's the big deal?

 

Idea number 2


 
Vinin:
512 min. =Uncle Kolya
Took it off the tongue :))
 
Vinin:

512 min =Uncle Kolya

I forgot to add that you can also use a fixed lot

Martingale Minutes - I've never seen anything like it.

1...24252627282930313233343536
New comment
Reason: