Avalanche 6.2 - page 2

New comment
 
SergNF:


And the point of your

is not clear.

Yes and the division into LotSell and LotBuy too.

LotLast and full stop.

About history analysis (MODE_HISTORY) - these are plans to make Avalanche logic on one order to reduce the deposit.
And division into LotSell and LotBuy is taken from Avalanche 6.2. I didn't change anything, just improved the code a bit.

 

wmlab:

The main problem with Avalanche, it seems to me, is the constant pushing of the break-even level further and further away from the starting point if orders are "reeled in".

I think you misunderstand the point of an Avalanche.

With the lot multiplier being equal to 2, the breakeven level is in place at a distance equal to the width of the open positions channel.

At a ratio greater than 2, the breakeven level is even closer to the channel of open positions.

If you propose the idea of not "winding up" the open positions, but leave the avalanche essence, that would be very interesting.

 
wmlab:

I've tweaked the 6.2 code a bit to work on the real. Here you go.

Fixed - version 0.3 dated 24.08.2010 (removed LoadHistory and MaxProfit logic). Why does it work "wrong"? What exactly is "wrong"?


On the tester periodically gives error 130 (wrong stops) and hangs on this point, I put printers, it turned out in OpderSend price=0 and it takes it from LevelBuy, can you look at the logic at your leisure?
 
wmlab:

I've tweaked the 6.2 code a bit to work on the real. Here you go.

Fixed - version 0.3 dated 24.08.2010 (removed LoadHistory and MaxProfit logic). Why does it work "wrong"? What exactly is "wrong"?


While you are re-inventing your wheel, a new version of Lavina_V63 has appeared. According to the test, it works very well and is stable. No errors have been detected.
 
OlegTs:
In the tester it occasionally generates error 130 (wrong stops) and hangs on this point, I put the prints, it turned out in OpderSend price=0 and it takes it from LevelBuy, can you look at the logic at your leisure?
I have written my version of avalanche on one single order. I will try it on real account today and will post it here or on codebase.
 
zhuki:
While you are re-inventing your bicycle, there is a new version of Lavina_V63. According to the test it works very well and is stable. No errors have been detected.


SO WHERE IS IT ?

Would love to see....

 
Galina:


SO WHERE IS IT ?

Would love to see....

I thought it was your joint brainchild with Rumata...
 
wmlab:

I've tweaked the 6.2 code a bit to work on the real. Here you go.

Fixed - version 0.3 dated 24.08.2010 (removed LoadHistory and MaxProfit logic). Why does it work "wrong"? What exactly is "wrong"?


Yes unfortunately there is an unavoidable loss. I think we should limit the time of its work. I put the time of installation of pending orders at 9.00 and the end of work with the closing of all orders at 22.00. The result is $ 3000 for half a year. A drawdown of 7% (depo 10,000). I am not calling for anything or arguing with anyone, just a fact.
 
143alex:
Yes unfortunately the inevitable drain is coming. I think we should limit the working time of the EA. I put time to set pending orders at 9 a.m. and close all orders at 10 p.m. The result is $ 3000 for half a year. The drawdown is 7% (depo 10,000). I am not calling for anything or arguing with anyone, just a fact.

I thought the problem was in the code =) Anyway - any martin will sooner or later lose everything. Filters will only slow down the growth of the balance and push back the inevitable drain. The trick is to withdraw your profits before the plunge comes.

 
wmlab:

I thought the problem was in the code =) Anyway - any martin will sooner or later lose everything. Filters will only slow the growth of the balance and push back the inevitable drain. The trick is to get your profits out before the flush comes.


Or test over long intervals (10 years, for example). With such a history of trades, with > 500 trades, one can hope that the system is suitable for real.
123456789...19
New comment