Building your own EA G/08.

[finimej]

Re: Building your own EA G/08.

 

have problem to download the final results, http://rapid*share.com/files/328262516/FxPro_-Real.rar

 

"Unfortunately right now our servers are overloaded and we have no more download slots left for non-members. Of course you can also try again later."

[asgard2]

Re: Building your own EA G/08.

 

 

You need to consider learning and sticking with manual trading systems.

 

Hmm, that's a wee bit presumptuous..LOL

 

I have been trading full time now for 2 years and I make just as many mistakes as my EA's. Only difference it that the EA's do it quicker.

 

Edit: Happy New Year to you also and may it be prosperous for us all here.

[Cyrillic]

Re: Building your own EA G/08.

 

Looks to me like all the do it yourself people quit once again :( Sad but true...

[hitescape]

Re: Building your own EA G/08.

 

Looks to me like all the do it yourself people quit once again :( Sad but true...

Well, I'm here to help if I can. I was looking at TradeChimp but their system is too $$$. It is a AudNzd scalper. If there is there anything I can do to help your version just let me know. PM is best.

[asgard2]

Re: Building your own EA G/08.

 

Looks to me like all the do it yourself people quit once again :( Sad but true...

 

LOL.. what do you mean?

 

I think we are all here... I am working on lots of ideas at the moment and I have some better results from the Developing EA.

 

If anyone cares I can post them but it takes a while to get something good.

 

If anyone wants to share their progress results, it would be great to see.

 

Somedays I think it will never work, otherdays I make a breakthrough and amaze myself at the improved results of a simple idea.

 

So every idea needs coding and testing and then if it works.. well, is it curve fitted?.. lots more checks to see if I am kidding myself.

 

Here is a couple of unusual ideas that improved performance and DD since I last posted.

 

When an order stops out, close all other orders of the same type or set breakeven.

 

When bear tick count is greater than 18 close buy orders. This can only be optimised of course on each tick. My PC is Dead.

 

Right now, I have just bought a new PC and I am waiting for it to arrive.. don't let anyone tell you you can't wear out your processor:)

 

I am also playing with modes and early warning systems.

 

Example below that is self explanatory.

 

if (MA4_P6_Close > MA4_P6_Open){
       if  (iOpen(NULL,PERIOD_M15,0) < iMA(NULL,0,iMA_BuyPeriod,0,MODE_EMA,PRICE_LOW,0) && iMA(NULL,0,iMA_BuyPeriod,0,MODE_EMA,PRICE_LOW,0) - PricePoint *Point > Bid ) OpenBuy=1; }

 

This works well but it is still four hours short of perfect.

[asgard2]

Re: Building your own EA G/08.

 

Just a note..

 

I shared this file with 4 others here, so far and before I deleted it, it had 18 downloads.

 

I have the "little red hen" thing going... Just because it is right and fair.

 

So if you have it, I don't mind and I did not ask those I gave it too, not to share it.

 

However if you care for a copy of the final, then please come up with some ideas or code as well.

 

I will get around to making a vid of this as soon as I get a pc capable of running "every tick" backtests again.

 

Right now DD is around 10% starting from Feb 09 though to Yesterday... with RD at around 48 something. Net Profit just barely exceeds the mill.

 

Still haven't started on the sell logic and still cannot get through Dec08 or January09 without degrading profitability unacceptably.

 

If it can get though DEC 08 with Buy enabled.. should about get through anything. :-?

[finimej]

Re: Building your own EA G/08.

 

the most important thing is

how to avoid to get caught on reversal little up when it is big down trend.

[asgard2]

Re: Building your own EA G/08.

 

the most important thing is

how to avoid to get caught on reversal little up when it is big down trend.

 

At the end of the day.. I think we will always have losses. The thing at the moment is to get entry and exit it as accurate as possible before we introduce the stoploss.

 

Try this code, it works like a dynamic stop and it still makes a fair profit.. I must pass on my candle rules, which are not in the code I gave you so far because I am still testing them for best profit.

 

OrderStoppedOut=false;

  if (MayOpenOrder){// Time Delay between buy stops or any closed order ........ //added the line about order stopped out
        total = OrdersHistoryTotal();
        for(i = total - 1; i >= 0; i--){
        OrderSelect(i, SELECT_BY_POS,MODE_HISTORY);
        if(OrderSymbol() != Symbol())  continue;  {
       if (OrderType() == OP_BUY && OrderClosePrice() - OrderOpenPrice() < TakeProfitB && OrderCloseTime() ==TimeCurrent()) bool OrderStoppedOut=True; 
if ((type==OP_BUY) && ((TimeCurrent() - OrderOpenTime())/60) < NextBuyStop_Order_Minutes) OpenBuy=0;
        if ((type==OP_SELL) && ((TimeCurrent() - OrderOpenTime())/60) < NextSellStop_Order_Minutes) OpenSell=0; 
          }} }
       
//added all of the below code.
if (OrderStoppedOut==True){ 
     
        total = OrdersTotal();
        for(i = total - 1; i >= 0; i--) {
        OrderSelect(i, SELECT_BY_POS, MODE_TRADES);
        if(OrderSymbol() == Symbol()) CloseOrder();
        }} 
          

 

That makes for an interesting visual but it would not be my preference ordinarily.

 

With your coding skills I am certain you could come up with a better solution. This is what I think but I have no idea yet how to make it happen.

 

If the second order is stopped out, close the first order with a higher open price in Just in case the market is going backwards. If it is going backwards it will simply buy at a lower price to replace the order it closed. If the market rebounds, it is the least profitable order to loose since the risk on the first order is the lowest risk.

[Cyrillic]

Re: Building your own EA G/08.

 

I have personally downloaded your file on a number workstations (i have over 5) So that might have added up on the count a little. Anyway I've been busy working on my main project (An artificial intelligence system designed to predict the future flow of the price action in a span of one week and less accuracy averages 79% so far) We've been trying to shrink the model to make it predict m15 and hourly data (main problem is time since a single calculation takes around 5 hours I can post the translated article about it) anyway, Im thinking of sharing the skeleton of aud/nzd scalper with some of the members of this forum if they want to work on improving it further since it does me no good lying around collecting dust. As for the system you send me i don't think i will have any time available to work on it but I will try and help if anyone has anything they would like me to add.

Anyway nice to see everyone still working hard and hope ya'll had great holidays.

Philippe.T

[fcorbier]

Re: Building your own EA G/08.

 

(An artificial intelligence system designed to predict the future flow of the price action in a span of one week and less accuracy averages 79% so far) We've been trying to shrink the model to make it predict m15 and hourly data (main problem is time since a single calculation takes around 5 hours I can post the translated article about it)

 

I'd be happy to read this article if you could post it. Even though I'm highly skeptical about anything that could "predict the future", I'm sure it'll be pretty interesting :)

[asgard2]

Re: Building your own EA G/08.

 

I'm highly skeptical about anything that could "predict the future", I'm sure it'll be pretty interesting :)

 

LOL.. Well if God and Santa are true, then I am skeptical of nothing and anything is possible :)

 

Just incase either of the above not true or not they are not helping.. I think we better stack the deck in our favour somewhat.

 

I would rather trade on the situation at hand right now, than the one that passed yesterday.

 

Meaning there are various things you can do with solid data that is coming in right now in the form of price, tick action, spread and last candles combined with trading style that will minimise risk.

 

For Example if the last day closed a dojie, you will know 1 second after it did, you will also know if the Market is High..This is data coming in this second, that will help you avoid a bad trade.

 

You can never predict a good trade, you can only avoid bad ones. Avoid enough bad ones and the good ones will be of a higher proportion. Risk is about so much more than percentage of your account vs stoploss level.

 

Trend is the biggest killer (along with the saying Trend is your friend)..who can predict where or when one will start or stop? These days I give little weight to trend. One thing is certain, anyone who bought at the end of a trend, bought way over price.

[asgard2]

Re: Building your own EA G/08.

 

(An artificial intelligence system designed to predict the future flow of the price action in a span of one week and less accuracy averages 79% so far) We've been trying to shrink the model to make it predict m15 and hourly data (main problem is time since a single calculation takes around 5 hours I can post the translated article about it)

Philippe.T

 

Hmm, 5 hours is a long time..surely a good trade entry has been and gone?

 

I calculate the next price this way.. and it if any future predicting type systems are good, then this will be about the same.

 

double Hour4MA8Shift=(iMA(NULL,240,5,0,MODE_EMA,PRICE_LOW,2)-iMA(NULL,240,5,0,MODE_EMA,PRICE_LOW,1)) / 0.0001; // How far did the MA drop between the last price and previous price
double PriceCalc = iMA(NULL,240,3,0,MODE_EMA,PRICE_LOW,1)- Hour4MA8Shift * 0.0001; //take that drop and apply it to the current angle to find a calculated next price. 

 

If you set it up, correctly you can optimise it to find the best combination for your pair. The idea is that it should filter out the market noise and minor retracements

[pokeregis01]

Re: Building your own EA G/08.

 

Right now, I am modifying another really good EA called MasterMind3, I found it here originally disguised under the name of Fred. It is the championship edition.. I believe this will be another 200,000 a quarter EA when I am finished.

 

Have you finished the modifications with MasterMind3?

What were the results.

[asgard2]

Re: Building your own EA G/08.

 

 

Have you finished the modifications with MasterMind3?

What were the results.

 

Actually I never finished it..

 

I started on something else that seems to be working, so I will stick with it.

[IamJon]

Re: Building your own EA G/08.

 

Thoughts on SellLimit and BuyLimit orders.

 

So the purpose of the placement of the Limit orders is that the expectation that a reactionary point where a price bounce (reversal) is likely to occur at price levels just beyond the SellLimit or BuyLimit values. These reactionary levels would be where expected support or resistance is. Some currency pairs may have regular/fixed price levels where support/resistance may occur. Some say that the GPBUSD regularly bounces at the "figure" + 20pips and the "figure" + 80pips. One would have to do some analysis on a given pair to see if there are sweet spots. In the Greezly code, we can see that the Acceleration indicators is used to delete the Sell or Buy limit orders depending on the sign of the indicator value. The deletion of limit orders based in the Speed indicator is a little more complex. More on this to follow.