Anti-Martingale vs Martingale : Good prevails!!! - page 11

 
MetaDriver >>:

Подумаю :)

Это не проблема - кредиты в казино беспроцентные. :)

And you're Brutus, neglecting swaps!
;)
I should add. And the shoulder, too.
Kolya Morzhov should come.

 
avatara >>:

...............

Коля Моржов приходить должен.

Uh-huh. A man has to be mortal. Otherwise, it's not a man!

 
MetaDriver писал(а) >>

I almost agree.

There are two subtleties:

  1. even this toy demonstrates that at MO>0.5 martin loses to anti-martin.
  2. when the average continuous series of winning/losing trades is lengthened (which is typical for meaningful strategies), the anti-martin becomes really profitable, even compared to a constant lot.

With constant positive MO there is an optimal share of capital (optimal fi) that should be set to achieve the maximum profit/risk.
Trading with a fixed share sets the anti-martin - when the capital increases (a series of profitable trades) the lot increases, when it loses - it decreases. In short, if there is a positive MO and it is invariable (which in reality is unlikely), then the optimal strategy would be to bet a certain fixed proportion of capital in each game. And one can not be guided by previous wins/losses as Martin/Antimartin does - only by the current amount of capital.
 
Avals >>:И можно не ориентироваться на предыдущие выигрыши/проигрыши - только на текущую величину капитала.

And you can navigate and get the best optimal capital share. True, neither martin nor anti-martin apply to this opera.

 
TheXpert писал(а) >>

And you can navigate and get the best optimal capital share. True, neither martin nor anti-martin apply to this opera.


This is already about seriality. It's really out of the question here. It was modelled without it as I understood it.
If a series of negative trades increases the probability of a positive one, then martin makes sense. If a series of positive trades increases the probability of a positive one, then anti-martin makes sense. But this is beyond the scope of the proposed simulation.

 
Avals >>:


это уже о серийности. Здесь о ней действительно речи не было. Моделировалось без нее как я понял.
Если при серии отрицательных сделок увеличивается вероятность положительной, то мартин имеет смысл. Если при серии положительных растет вероятность положительной, то антимартин. Но это за рамками предложенного моделирования.

Yes, absolutely right. So far outside the box.

I'm in the middle of a series now.

Made a series length calculation - statistics on the fact of obtained random series. // Taking in the trailer.

I think, if I make adjustable series (it's not hard, I've already thought of such a generator), it could be something useful. It would be possible, say, to load there rows of transactions from the tester and select (optimize) MoneyManagement.

But I'm sick of this Excell, with its tardiness. The main time is consumed by the output of results in squares.

I cannot avoid it, because the charts are plotted in cells.

Why don't I make such a thing in Delph, or even better in Sharp? To make such a YopenSource project (probably already on mql5.com). Like MoneyManagement Studio :).

And build up slowly. I think ... I estimate the ratio of motivation / laziness. :)

Files:
 
avatara >>:

А ты Брут, пренебрегавший свопами!
;)
Надо добавить. И плечо тож.


Maybe it's in the next project. :)

As long as the modelling is of the "casino" type - the size of the win/loss depends only on the size of the bet.
// Here's the commission added - I'm asking you to include everything there in bulk! :)
--

Svinozavr 25.04.2010 15:17
avatara >>:
...............

Kolya Morzhov should come.

Yes, a human being must be mortal. Otherwise he is not a human being!
And I can call Morzhov, no problem. // He's always there. Behind his left shoulder.
In the next version, especially for humanists...
;)
 
MetaDriver >>:

Возможно эт уже в следующем проекте. :)

Пока моделирование "казиношного" типа - размер выигрыша/проигрыша зависит только от размера ставки.
// Вот комиссию добавил же - туда прошу всё и включать оптом! :)
-- А Моржова могу позвать, без проблем. // Он всегда тута. За левым плечом.
В следующей же версии, специально для гуманистов...
;)

MegaProject will be, too. ;)

Swaps are interesting because they are different. Sometimes they're on the plus side...

And leverage of 500 gives unheard of returns.

 
MetaDriver >>:
..........

Мож наваять такую штуку на Делфе, а ещё луче на Шарпе? Сделать такой ЁпенСоурс проект (наверное уже на mql5.com). Типа MoneyManagement Studio :)

И наращивать потихоньку. Думаю... прикидываю соотношение мотивации/лени. :)

Why bother with Delphi and Sharpies? You can use MQL5, thank God everything for such a program is available - custom buttons and other self-made controls can be easily created.

And I'd like the motivation / laziness to be higher than 1, if only the MQL5 program skeleton would be created.

 
MetaDriver >>:

...

Мож наваять такую штуку на Делфе, а ещё луче на Шарпе? Сделать такой ЁпенСоурс проект (наверное уже на mql5.com). Типа MoneyManagement Studio :)

И наращивать потихоньку. Думаю... прикидываю соотношение мотивации/лени. :)


If you want to go so global, you should definitely write in Sharp. Here you have both LINQ and WPF, in general, compared to this, all the rest are yesterday's day. By the way, I'm currently reading a book on C# - I'm in raptures of delight. And how did they program without it before?