Simple 3 order GBPUSD EA

Is the lot size fixed or on Money Management?

It's on MM. I remember it's on 1.5% per order.

I made a backtest from 2007 to now and its how it looks...

Do you account for a possible pip widening at the weekends ? Most brokers have much wider pips ! Your results of backtesting could be pretty distorted !

I really don't know. I'm using the deftual settings with alpariuk

Hi Geoff,

First of all, great job in developing this amazing strategy and EA . I'm always sceptical when I see EA results like those you show on page 1 but, sure enough, I tried your EA and it works great .

If you don't mind, I have a few suggestions on how to polish up the coding of the EA. The polish won't effect the great results in anyway but should make the EA more efficient and easier to maintain going forward. I'm a computer programmer by profession but nothing to do with FX or the financial markets. However, I have been interested in Technical Analysis for 10 years and I've been dabbling with EAs for a while now. I would be happy to help.

So if you don't mind here are just a few suggestions having looked through camvcvooV5.mq4 (I'd be happy to make these changes myself and send them yo you):

1) The code that works out the average and current daily range should be enclosed in an if(Show_Settings) { ... } so that this computation is only done if really needed - working this out on every tick when you don't need it will waste CPU.

2) The 15min Fractal code that works out TimeOfLastUpFractal_H1 and TimeOfLastDownFractal_H1 is currently being called at every tick. This should be moved down to where the Place Order code is after it checks Hour() == trade_hour so that the 15min fractals are only queried when needed. This should speed things up a bit more.

3) Is the code under the Close Orders comment still needed or was this only needed when there was an option to enter positions without a SL and TP ?

4) In a number of places there are 2 calls to GetLastError() to check if an OrderSend/OrderModify was successful or not. The first checks for the success of the Order and the second prints out the error number if it wasn't. The problem with this is that the first call to GetLastError() resets the error number so the second call will always return 0. I'm sure reading through this thread a number of people have been getting Error # 0 messages and could not figure out what was wrong.

To remedy this you should use the return value from OrderSend/OrderModify to determine success or failure of the command. OrderSend returns the ticket number of the order or -1 if it fails and OrderModify returns true or false depending on whether the order was modified successfully or not. So I suggest the following style:

int ordTicket = OrderSend(...);

if(ordTicket == -1) {
Print("Buy OrderSend 1 failed with error #", GetLastError());
sleep(1000);
}

bool modOK = OrderModify(...);

if(!modOK) {
Print("Buy Order 3 Modify Failed With Error #", GetLastError());
sleep(1000);
}

5) In V5 I see the OrderSend has been split into an OrderSend (without SL and TP) followed by an OrderModify (with SL and TP) but I think there could be an issue here because the OrderModify may fail for some reason and it will go back around the retry loop issuing multiple OrderSends until the OrderModify succeeds.

6) There is a lot of repeated code for handling each of the 3 orders. For example you could create a sendOrder(int orderNumber, double takeProfit) function that contains the common code and just call it 3 times:

sendOrder(1, TakeProfit1);
sendOrder(2, TakeProfit2);
sendOrder(3, TakeProfit3);

int sendOrder(int orderNumber, double takeProfit) {

int retries =10;
while(retries > 0) {
RefreshRates();
int iTicket=OrderSend(Symbol(),OP_BUY,dLots,Ask,3*Poin,0,0,trade_comment,iMagicNumber,0,Blue);
bool xbool = OrderSelect(iTicket, SELECT_BY_TICKET);
OrderModify(iTicket,0,OrderOpenPrice()-Poin*StopLoss,OrderOpenPrice()+Poin*takeProfit ,0,CLR_NONE);
//OrderSend(Symbol(), OP_BUY, dLots, Ask, 3*Poin/Point, Ask - StopLoss * Poin, Ask + TakeProfit1 * Poin, trade_comment, iMagicNumber, 0, Blue);
if(GetLastError() < 0)
{
Print("Buy OrderSend ", orderNumber, " failed with error #", GetLastError());
Sleep(1000);
}
else
{one_shot=0;
break;}
retries--;
}//end while

}

Regards,
Piptician

You guys gonna fire up the ea with US holiday tomorrow?

AFFIRMATIVE!

Yes, I'm going to leave geoffs wonder on. I want to test it deeply even on extreme conditions.

OF COURSE !!

This is NFP week, so lots of movement to be expected.

ok

Victor, good morning to you!

Did you ever think about the following:

Backtesting v7a 2 leg strategy vs 3 leg strategy brings the following results:

v7a has roughly max drawdown of 55% of 3 leg strategy but produces 65% of profit if backtested till 1st of Jan 2010. Hence profitability probability is better if you double up v7a instead of using 3leg strategy.

Make up your mind and tell me what you think plse? Would be really interested because I am thinking towards increasing my risk profile and wonder if this would be a better way to facilitate......



cheers
roman

This is great, if you don't mind I will PM you regarding your suggestions!

EDIT: Can you enable PM's in your profile please?

I've just enabled email contact by members but it turns out as a Junior Member I can't send private messages or emails to other members. Looks like you can
though.

Hi Roman, honestly speaking, i really don't know. I have not done any backtest on any of the two EA. I just use them live forward until either one breaks down (hopefully never!).

To me, the bet is for GU move +22 pips in our favour. Once reach, everything else is bonus. If that the case, with 3 legs, i get minimum +22 and +10. With 2 legs, i only get +22 minimum.

Geoff, what do you think?

Regards,

Victor V

I am very impressed with the results we have had for this EA and I thank you for coming up with the idea.

For me, 3 orders is a bit of a problem because I tend to work on a much lower risk than many people,
even 2 consecutive orders is high risk for me!

The ideal EA for me would place a single order, make a profit and get out within the hour, anything that
runs for longer than an hour is too stressfull!

That's why manual trading is no good for me!

no orders for me last night. I turned on maybe 30 minutes after market open. Is that a prob?

No, I don't think anyone got a trade, you can turn on right up to the start time with no problems!

cool geoff, thanks.

I should have explained my last post a bit better!

The EA does nothing all day, that's the life, then at the start time it looks back
to the two previous hours for the "trend" if they agree it then looks to the M15
and M5 fractals for a suitable entry point.

Of course, if the two SMA's do not agree, it takes the next day off as well!