Avalanche - page 192

 
artikul 19.04.2010 09:29

mig34 wrote (a) >>

avalanche to slowly but surely make a profit


What does "avalanche" have to do with it? Any open position within a few days can produce a profit or loss several times the size of the initial margin. But it is unclear, if everything is so cloudless, why have not you screwed your own wisdom to the real? Or you do not have enough guts to make a profit after 9 plums? I'm sorry, what reason would you have for those who read your lies and believe them - to lose money steadily? You'd better tell the truth about the kitchen where you and John come from and what you're broadcasting here. )))

You are bullshitting because you are greedy for money
Today you made a profit after a flat, read the avalanche (for an exit) dub
 
rumata1984 писал(а) >>

What's wrong with my version that I posted here? )

I understand the "girl talk", but the first order (only the first order) with a takeprofit spoils the whole picture.
From your own report:

27 2010.01.05 19:48 sell 11 0.01 1.43564 0.00000 1.43164
28 2010.01.05 19:48 buy stop 12 0.02 1.44181 0.00000 0.00000
29 2010.01.05 19:48 modify 11 0.01 1.43564 0.00000 0.00000
30 2010.01.06 07:46 delete 12 0.02 1.44181 0.00000 0.00000
31 2010.01.06 08:56 close 11 0.01 1.43305 0.00000 0.00000 2.58 316.72
32 2010.01.06 08:56 sell 13 0.01 1.43290 0.00000 1.42890
33 2010.01.06 08:56 buy stop 14 0.02 1.43907 0.00000 0.00000
34 2010.01.06 08:56 modify 13 0.01 1.43290 0.00000 0.00000
35 2010.01.06 16:08 buy 14 0.02 1.43907 0.00000 0.00000
36 2010.01.06 16:08 sell stop 15 0.04 1.43290 0.00000 0.00000
37 2010.01.07 12:46 sell 15 0.04 1.43290 0.00000 0.00000
38 2010.01.07 12:46 buy stop 16 0.08 1.43907 0.00000 0.00000
39 2010.01.08 14:09 delete 16 0.08 1.43907 0.00000 0.00000
40 2010.01.08 14:09 close 14 0.02 1.42712 0.00000 0.00000 -23.98 292.74
41 2010.01.08 14:21 close 15 0.04 1.42736 0.00000 0.00000 22.13 314.87
42 2010.01.08 14:21 close 13 0.01 1.42736 0.00000 0.00000 5.51 320.38

'
The order which actually 'played the chain' is number 15, which is SELL, while the order (SELL) that could have played it, number 11, was closed ahead of time.
In other chains the first orders turn out to be opposite to the winner, but . truth (algorithm) is more expensive.
If you have to 'fear uncontrollability', then takeprofit all orders and modify TP each time a new position is added.
 
sever29 >>:


рост депо лучше, чем у меня. у тебя одна нога, у меня другая, однакож ходить надо на обеих:)

Of course you are! I was by no means thinking of competing with you. On the contrary, I'm very glad that someone else is doing it. Maybe you can suggest something interesting. In turn, I can suggest (although it has already been written here, but in case you haven't noticed), that a virtual trailing stop works great.

 
lexandros >>:
Вот стейты с реала например каждую пятницу - были бы убедительней любых самых бредовых расчетов... Вот он стейт ребята - нате еште... На понедельник было 10000 - на пятницу закрылся в 20000 - вывел 10000..
Вот это аргумент, с которым не поспоришь...
This is not an argument. Do you want to see the weekly state of a super-trader on a demo account, who will not make a mistake in the direction of price movement once in a week, and will close every day with a profit, having placed only one order in the morning? It is very simple: you open 32 demo accounts and every day you open a different order - either Buy or Sell. On the first day the first 16 accounts will be occupied by Buy, and the other 16 will be occupied by Sell. Of course, only half of the accounts will be profitable, those which will not be profitable are discarded. On the second day, we again open Buy on 8 of the remaining 16 accounts and Sell on 8. Only a half of them will guess, the rest will be discarded. And so on. At the end of the week you will have one stand of a super-trader, who was winning all five days of the week!
Next week I will repeat this trick, unobtrusively removing the account number from the screenshot (explaining that I don't want to share such personal information with anyone). I could do this every week indefinitely. What will you do? Will you start praying for me?
So don't expect any stats - they are all easily forged and tell you nothing.
 
SergNF >>:


Я понимаю, что "девушка уговорила", но первый ордер (только первый ордер) с тейкпрофитом портит всю картину.
Из Вашего же отчета:

27 2010.01.05 19:48 sell 11 0.01 1.43564 0.00000 1.43164
28 2010.01.05 19:48 buy stop 12 0.02 1.44181 0.00000 0.00000
29 2010.01.05 19:48 modify 11 0.01 1.43564 0.00000 0.00000
30 2010.01.06 07:46 delete 12 0.02 1.44181 0.00000 0.00000
31 2010.01.06 08:56 close 11 0.01 1.43305 0.00000 0.00000 2.58 316.72
32 2010.01.06 08:56 sell 13 0.01 1.43290 0.00000 1.42890
33 2010.01.06 08:56 buy stop 14 0.02 1.43907 0.00000 0.00000
34 2010.01.06 08:56 modify 13 0.01 1.43290 0.00000 0.00000
35 2010.01.06 16:08 buy 14 0.02 1.43907 0.00000 0.00000
36 2010.01.06 16:08 sell stop 15 0.04 1.43290 0.00000 0.00000
37 2010.01.07 12:46 sell 15 0.04 1.43290 0.00000 0.00000
38 2010.01.07 12:46 buy stop 16 0.08 1.43907 0.00000 0.00000
39 2010.01.08 14:09 delete 16 0.08 1.43907 0.00000 0.00000
40 2010.01.08 14:09 close 14 0.02 1.42712 0.00000 0.00000 -23.98 292.74
41 2010.01.08 14:21 close 15 0.04 1.42736 0.00000 0.00000 22.13 314.87
42 2010.01.08 14:21 close 13 0.01 1.42736 0.00000 0.00000 5.51 320.38

'
Ордер, который собственно "сыграл цепочку" - номер 15, который SELL, а ордер (SELL), который мог бы подыграть номер 11 Вы закрыли раньше времени.
В других цепочках первые ордера оказываются в первые ордера оказываются противоположными к победителю, но ... истина (алгоритм) дороже.
Если уж "бояться неуправляемости советника", то тейкпрофить надо все ордера и модифицировать TP при добавлении каждой новой позиции.

This is correct. In the following versions, after the take has failed and doubling has started, the take profit is removed. And there is a version without it at all. I have made a lot of them). For all tastes. And if you want to see another unexpected one, I can make it too.

 
rumata1984 писал(а) >>

This is correct. In the following versions, after the take has failed and doubling has started, the take profit is removed. And there is a version without it at all. I have made a lot of them). For all tastes. And if you want to see another unexpected one, I can make it too.


Try screwing in what I suggested... A variable corridor, as suggested by SAM.

 
sever29 >>:

моя версия лавины, в тестере, не сливает, хоть ты тресни. :) Любой временной интервал с одним параметром.

And it won't. That's the idea. "Avalanche is the only break-even system.

 
rumata1984 писал(а) >>

Of course you are! I was by no means thinking of competing with you. On the contrary, I'm very glad that someone else is doing it. Maybe you can suggest something interesting. In turn, I can suggest (although it has already been written here, but in case you haven't noticed), that a virtual trailing stop works great.


space... How is it different from a simple one, apart from the fact that it is not seen by the DC (if I understand correctly)

 
JonKatana писал(а) >>

And it won't. That's the idea. "Avalanche is the only break-even system.


and you with humor.

 
rumata1984 писал(а) >>

And if you want to see some more unexpected ones, I can do that too.


I've made a lot of "netting" and "locked" versions myself :)
Last dabble - lot increment and change of channel width depending on iteration number.
But I'm not interested in a graph of anything over a few hundred years from a few meters, but in a fragment of the graph in front of the maximum lot.
Here I wanted to derive a formula (most importantly, a nice one!!!) for the "equivalent takeprofit" for each iteration.
I.e. we have the second order with lot X to take a profit of Y, we should set TP, and SL for the first order; the third order has opened, etc. And we have any lots of each new order and any trigger level. So to say, to build a chart of "spread legs of the circular" (I don't remember all terms long ago).
But ... my idle time is over :(
'
SZY... There is also a version when a chain reaches the specified profit (still in "locked system" all is easier to count) all orders of unprofitable direction are closed, and the orders of "profitable direction are trailed", but ... multiplying virtual capitals is bullshit (about 7 variants are already standing on the demo). It is more interesting (was) to try to cheat ... "world behind the scenes" :) on unpleasant plots.
(I was better at training grids to predict volatility changes than anything else :) )