100% Win Math Grid Ea

Quoting wwwild

Disliked

Hi Viper, this is a modification I am using so that the EA will open up "full" lot sizes.

i.e. instead of opening 6 pending orders @ 0.01 lot size it will open 1 pending order @ 0.06 lot size.

This has a few main benefits -
1. Massively reduces the number of pending orders.
2. Makes the EA slightly quicker
3. When viewing backtesting results, I can easily sort by lot size to see the largest lot size that was pending and the largest that it opened. Without this mod, this info is virtually impossible to obtain.

You will see all of the...

Ignored

Quote

Disliked

if(MONEY_MANAGEMENT) LOTS=NormalizeDouble(AccountBalance()*AccountLeverage()/1000000*RISK_RATIO,0)*MarketInfo(Symbol(),MODE_MINLOT);

Quote

Disliked

if(MONEY_MANAGEMENT) LOTS=NormalizeDouble(AccountBalance()*AccountLeverage()/10000000*RISK_RATIO,0)*MarketInfo(Symbol(),MODE_MINLOT);

Quoting wwwild

Disliked

Hi Viper,

When backtesting I noticed that the EA gets slower and slower and slower... In fact to backtest a year of data was taking days. After looking at the code, I saw the reason why.

The function haveClosedOrdersFromSet is called every tick.

This function goes through every order in history. Hence, when backtesting, if you get up to 15,000 closed orders, every single tick the EA iterates through every one of them, hence the EA gets slower and slower and slower.

It might be an idea to add a counter in there that increments...

Ignored

Quoting Heikken

Disliked

Hello all...

I change a little about the code of money management of the first original EA.
from :

to


Then I run it with parameters:

Increment : 15
Level : 3
money management : true
risk ratio : 2
balance :5000

From 1 january 2010 till 31 January 2010 it became 8116.45


I got almost double balance in every month (Jan, feb, March) I backtested

But every month I always find at least 1 drop balance because of 1 loss session.

Then when I did backtesting from 2010 January 01 to 2010 April 20.

I stopped it when it still running in...

Ignored

Quoting wwwild

Disliked

Hi Viper, this is a modification I am using so that the EA will open up "full" lot sizes.

i.e. instead of opening 6 pending orders @ 0.01 lot size it will open 1 pending order @ 0.06 lot size.

This has a few main benefits -
1. Massively reduces the number of pending orders.
2. Makes the EA slightly quicker
3. When viewing backtesting results, I can easily sort by lot size to see the largest lot size that was pending and the largest that it opened. Without this mod, this info is virtually impossible to obtain.

You will see all of the modifications...

Ignored

Quoting wwwild

Disliked

Hi Viper, this is a modification I am using so that the EA will open up "full" lot sizes.

i.e. instead of opening 6 pending orders @ 0.01 lot size it will open 1 pending order @ 0.06 lot size.

Ignored

Quoting wwwild

Disliked

Hi Viper,

While investigating some unusual "losses" during backtesting, I found this bug (that caused the wierd losses) with an easy fix.

Basically, this bug will not come up very often, but is definatley there. What happens is the current code will only open 1 new pending order every tick. So if it had to open 60 pending orders, it would take 60 ticks. If the price reversed the opposite direction and hit one of the levels before 60 ticks, it would mean that it would not have enough pending orders to open in order for the maths to work correctly.......

Ignored

Quoting wcg

Disliked

Hi www, do you mind sharing the modification you did for speeding up the back test?, and where to put it.
Thanks
WCG

Ignored

//+------------------------------------------------------------------+
//| Function:  haveClosedOrdersFromSet
//+------------------------------------------------------------------+
bool 
haveClosedOrdersFromSet(string commentKey)
{
   if(Verbose && DebugPrint) Print("haveClosedOrdersFromSet; input commentKey="+commentKey);

   // If we have closed any orders from the set, close them all
   int totalOldOrders = OrdersHistoryTotal();
   
   //EJAZ We use variable to track how many past orders we have looked at
   int lookedAt = 0;
   
   for(int cpt=totalOldOrders-1; cpt >= 0; cpt--)
   {
      OrderSelect(cpt,SELECT_BY_POS,MODE_HISTORY);
      if(OrderSymbol()==Symbol() && OrderMagicNumber()==MagicNumber)
      {
         // For closed orders, MQ4 appends their own "close reason" comments to the end of yours. 
         // Use only your comment portion for comparison
         string closedOrderCommentKey = StringSubstr(OrderComment(),0,StringLen(commentKey));
         if(Verbose && DebugPrint) Print("haveClosedOrdersFromSet; closed order commentKey="+closedOrderCommentKey);

         if(closedOrderCommentKey==commentKey)
         {
            DebugPrint = false;
            return(true);
         }
         
         //EJAZ If we have gone past 40 orders, we dont need to look any further
         //EJAZ This speeds up backtesting considerably.
         lookedAt = lookedAt+1;
         if (lookedAt > 40) return(false);
      }
   }
   
   return (false);
} 

Quoting Heikken

Disliked

Hello all...

I change a little about the code of money management of the first original EA.
from :

to


Then I run it with parameters:

Increment : 15
Level : 3
money management : true
risk ratio : 2
balance :5000

From 1 january 2010 till 31 January 2010 it became 8116.45


I got almost double balance in every month (Jan, feb, March) I backtested

But every month I always find at least 1 drop balance because of 1 loss session.

Then when I did backtesting from 2010 January 01 to 2010 April 20.

I stopped it when it still...

Ignored

Quoting wwwild

Disliked

Sure, here it is. There are only 3 lines of code. To see which are the new lines, simply have a look at the //EJAZ comments.

If you want to use this, simply find this function in the EA and replace it with the version below.

For those who don't know what this is for, please see this posting. Basically, this modification speeds up backtesting.

[php]//+------------------------------------------------------------------+
//| Function:...

Ignored

Quoting rbrtsha

Disliked

Quoting Heikken

Disliked

Hello all...

I change a little about the code of money management of the first original EA.
from :

to


Then I run it with parameters:

Increment : 15
Level : 3
money management : true
risk ratio : 2
balance :5000

From 1 january 2010 till 31 January 2010 it became 8116.45


I got almost double balance in every month (Jan, feb, March) I backtested

But every month I always find at least 1 drop balance because of 1 loss session.

Then when I did backtesting from 2010 January 01 to 2010 April 20.

I...

Ignored

how is it closing losses before margin being called?

Ignored

Quoting wwwild

Disliked

Hi Heikken, can you try applying this modification then run your test again?

It may not be related but I had similar issues (originally) during certain backtests. I managed to track it down to the way the EA opened pending orders. See if you can apply the modification linked above and see if that makes a difference.

Cheers, Andrew.

Ignored

Quoting Heikken

Disliked

Thanks Andrew...

Which line should I put these codes on?

(sorry I'm not a programmer)

Ignored

Quoting reachjj

Disliked

Why this thread is dead?... Can anybody point me to working version of this EA.

Ignored

Quoting NuCom

Disliked

After read from 1st thread till present, I found it very interesting.

My real account got some positive pips but when the ranging very high, it took it all.

I need helps from proggrammer here to modify a free EA., It will stand in rangging and trending... but reqoute is the worse scenario that can drag the losses order far away from the targetted Stoploss. It use stealth mode for it stoploss and takeprofit.

Please comment if somebody willing to help modify the EA.

Thanks.

Ignored

Quoting FrozenAir

Disliked

I have been editing your code Viper to create a more dynamic high-water mark scenario. I like what you have done so far. Even running the high-water mark with a position of 10% margin to begin engaging has netted me profit over my test runs. Obviously at 10% it got triggered a few times and took some profit from the bottom line, but it prevented any death trades and kept the trading going in the positive direction. I am making some modifications to deliver a positive result from the high-water mark being engaged. I will post the results after some...

Ignored