Once again, about the lokas. - page 19

New comment
 
avatara писал(а) >>

To the free will, and to the saved Heaven.

Good luck with that. ;)


>> Thank you. Your wish is the answer to all the panelists.

 
avatara >>:

;) Спасибо!

Формально Вы правы. Но свопами в торговле обычно пренебрегают.

Или Вы кэррикеш стратегии используете?


No, I don't - just formal proof.

Good luck.

It's true that Islamic accounts won't be affected, but do a lot of people trade in them? And if you ignore swaps, there is no difference.
 
lexandros писал(а) >>


Absolutely agree on one thing... That the market is much more often in a flat than in a trend. But no one can predict when there is a flat and when there is a trend... (at least until now)... Maybe on the next tick it will go in a 1 000 point trend... or maybe it will be flat for a couple of weeks.

And this is the essence of both of these evils... averaging as well as martin...
Or rather, averaging is not such an evil... I often use it myself... And sometimes the averaging gives a good result... If you use the averaging thoughtlessly and without thinking - then it is no better than martin...

Well, speaking of horses... I've written MTCs using both averages and martins - both in combinations... For myself and as an order...
If it works on a flat, then the instantaneous release of one of these signals may cause some disturbance. If it works on the flat, it will instantly lose on a strong trend and vice versa... I think everyone knows it.

That's why it's evil... Because any such MTS (or hand trading) - it will definitely sell out sooner or later, with 110% probability.
Combine both methods - no one has yet succeeded ... for if they succeed - it will be a grail.... (Maybe someone has succeeded, but he's sitting quietly somewhere on his own island, making trillions, and won't tell anyone the secret).

That is why I do not consider loki "evil incarnate". sometimes, i repeat sometimes - you can lock yourself in. When you're not sure it's time to close, when you're hesitating. And only with your hands... No robot will assess the situation correctly... (or vice versa wrong). You tell it to lock, it locks... ...and it stalls in a deep drawdown. And that's the end of the lamb. That's why MTS based exactly on locks is an inherent evil. Not the lots themselves, but MTS based on them. Because it's guaranteed to sell out.
Absolutely sure about it, until someone proves otherwise. i.e. show profitable MTS based on lots (without billions of initial deposits and paltry profits) which won't lose at least 5 years of history.


When they want an EA to be profitable for at least 5 years - this smells like childish maximalism. I think the Expert Advisor must be profitable for at least half a year - and that's good.
If it is not profitable, switch it to a demo account and wait. On the real one, put another one that has shown good results on the demo in recent times.... and so on to infinity.
 
lexandros >>:
Скомбинировать оба метода - пока еще никому не удавалось... ибо если удасться - это будет грааль.... (может и удалось кому-то, но он тихо сидит где нибудь на собственном острове и стрижет триллионы, и никому секрета не расскажет).
And rightly so.
For if he says anything, his method of distinguishing between a flat and a trend will stop working almost immediately. It may well happen all the time - someone finds a method to distinguish, as long as he keeps quiet - the method works, as soon as he says something (or starts showing off by having one with hints of criteria) - the criterion is immediately (almost) randomized. I'm sure - many bankers are lurking on forums looking for workable ideas. They don't need a lot of hints - the pot is brewing properly.
C'est la forex.
 
SProgrammer >>:

При случайном входе ( покупка или продажа и время ) вероятность срабатывания стопа или тейка будет ПРОПОРЦИОНАЛЬНА их размеру. То есть если TP = 20 а SL = 20 то веротность закрытия в прибыли будет равна верятности закрытию с убытком. Не зависимо от тренда и валютной пары и времени в истории. Ну а если TP = 2* SL то вероятность убытка будет в два раза выше. :)

Decided to check the statement in the tester. The Expert Advisor counts the number of ZigZag knees (at least Pips) and writes it to the file:

extern int Pips = 100;

int Amount;

int GetZigZagCount()
{
  static bool FlagUP = TRUE;
  static int Min = 999999;
  static int Max = 0;
  static int Count = 0;

  int PriceHigh = High[1] / Point + 0.1;
  int PriceLow = Low[1] / Point + 0.1;
  
  if (FlagUP)
  {
    if (PriceHigh > Max)
      Max = PriceHigh;
    else if (Max - PriceLow >= Pips)
    {
      FlagUP = FALSE;
      Min = PriceLow;
      Count++;
    }
  }
  else // (FlagUP == FALSE)
  {
    if (PriceLow < Min)
      Min = PriceLow;
    else if (PriceHigh - Min >= Pips)
    {
      FlagUP = TRUE;
      Max = PriceHigh;
      Count++;
    }
  }
  
  return(Count);
}

void StringToFile( string FileName, string Str )
{
  int handle;
  
  handle = FileOpen(FileName, FILE_READ|FILE_WRITE);
  FileSeek(handle, 0, SEEK_END);

  FileWrite(handle, Str);

  FileClose(handle);

  return;
}

void deinit()
{
  StringToFile("Analyse.prn", Pips + " " + Amount);
  
  return;
}

void start()
{
  static int PrevTime = 0;
  
  if (PrevTime == Time[0])
    return;
    
  Amount = GetZigZagCount();
  
  return; 
}
Runs in tester with optimization:
ׂ

Plots of the dependence of the number of knees on their min. size(Pips):
ׂ

Plots of TP and SL probability ratio(p(SL) / p(TP)) for TP / SL = Koef from SL size
ׂ
ׂ
ׂ
In the above graphs, the blue line is the Koef^2 value.

From the results, it appears that the original statement is not quite true. The following one is closer to the truth:
At random entry ( buy or sell and time) probability of stop or take will be PROPORATORIAL tothe square of their relation. I.e. if TP = 20 and SL = 20, the probability of closing in profit will be equal to the probability of closing with a loss. Regardless of the trend and currency pair and time history. If TP = 2* SL the probability of loss isfour times higher.


Hypothesis:

p(TP) / p(SL) == (SL / TP) ^ 2
 
getch >>:



Из полученных результатов выходит, что изначальное утверждение не совсем верно. Ближе к истине следующее:
При случайном входе ( покупка или продажа и время ) вероятность срабатывания стопа или тейка будет ПРОПОРЦИОНАЛЬНА квадрату их отношений. То есть если TP = 20 а SL = 20 то веротность закрытия в прибыли будет равна верятности закрытию с убытком. Не зависимо от тренда и валютной пары и времени в истории. Ну а если TP = 2* SL то вероятность убытка будет в четыре раза выше.


Something's not right here.
 
MetaDriver >>:
И правильно делает.
Ибо если сболтнёт - его метод различения флет/тренд работать перестанет почти сразу. Вполне возможно это и происходит постоянно - находит кто-то метод различения, пока молчит - метод работает, стоит сболтнуть (или начать выпендриваться наличием такового с намёками на критерии) - критерий сразу же (почти) рандомизируется. Уверен - многие банкиры пасутся на форумах в поисках рабочих идей. Большого количества намёков им не требуется - котелок варит исправно.
Се ля форекс.

Don't mess with people's heads.))) What are you kidding me for? Do you want people to develop delusions of grandeur in addition to the paranoia they already have? The market doesn't really care who uses what method. ===

(F-M backstabbing is off the table - shhhh))))

 
getch >>:

Решил проверить


i.e. if SL = 2* TP then the probability of closing in profit will befour times higher.
and in pips the difference is only 2 times
free grail turns out )

 
Mischek >>:

халявный грааль получается )

I see. The interpretation of the results is apparently wrong.
I attach MathCad file with data source.
With SL and TP there are doubts. But the following statement is exactly by interpretation:

Amount(Pips1) / Amount(Pips2) == (Pips2 / Pips1) ^ 2
where Amount(Pips) is the number of ZigZag knees with a knee of at least Pips
.

Files:
analysetmp.rar  2 kb
 
getch >>:

Понимаю. Интерпретация полученных результатов, видимо, ошибочная.
Прилагаю MathCad-файл с источником данных.


I believe it ) but it can't be.
I mean the result should be two (not four)
1...121314151617181920212223242526...61
New comment
Reason: