Opinion - very successfull EA - $3000 account to $6300 in two weeks (could have been $9000) - page 3

....and I will be giving the EA away when it works :-) dont worry....

Ok well, since you're going to be giving the EA away anyways. Why don't you provide the details within this thread. I'll be willing to code out the development cycle within this thread. Who knows, maybe some other super-coder would do it faster. The reason I'm offering this is because I'm currently motivated doing something for someone else. And because, I have the feeling you're going to keep posting how well your semi-automated system is doing. Lets code/ back-test/ forward-test/ penny-test and move on.

This is mql4.forum. Providing results without showing codes is pointless.

@ubzen - yes but it is his project and may want the feeling of having done it his way and enjoy the feeling that gives.

@ MickGlancy - just listing an option not actaully speaking for you.

sory that i am in the midle again

there is an old stradegy we had use since late 80's and it still works.

and because allways my motive before an entry is not how much money i am going to make but how much money i am going to loose

i start allways with position size : (account size x risk%)/amount of pips at SL=lot size

50,100 sma, MACD (stantard settings) when price's close above(Below) 50 or 100 sma (which of them prices past at last) and MACD histobar cross (0) line within 5 histo bars not more buy(Sell)

SL at the botton of the previius 5 chart bars (entry bar incl. ) when TP=2 x the amount of SL (or 1 x amount SL) close 1/2 of the position and move sl at break even ( 1st profit and the rest of the trade for free) the remaining 1/2 position we close when prices recross 50,100sma(which is the nearest) thats it !!!

it better works in 1h chart and daily

offen in 1h chart you can be in the market for some days, and in daily chart for weeks

we had count an accuracy of 60% min which is good because the proffits overcome the losses with money managment

there are other older and forgoten trating systems they had been used the old days and still work well

remember: think fundamental and react technically

Ok, fine Mr. fundamentalist, because I'm in love with MACD 8)) I'll go ahead and attempt to program your classic method. This one better show +profit within the same period I last tested. If not, I'll dump my own random system in here and blog my development process.

Update:

Looks like the Old-Timer's system beat the crap out of my little sample period. Horay lol. Here's his back-test from 1-2010 to 3-2010. That period usually have a mix of trend and ranges and this system handled it well. I've attached the EA and files again. The money-management and order close could definitely use some work. I was getting tired so that section is not as great as it could be. I intend to follow up with this system and provide extended statistical analysis for my own learning. Others can definitely modify the codes and make requests, and I'll see if we could include it.

Warning: This is not a complete EA, please don't run this thing on a real account.

Ive worked out how to draw data to the bottom window, Im going to try and draw a custom RSI to that window with the variables I put into extern int in the EA. thats should work eh ?

Be careful with your drawing processes. One cannot utilize most mql4 draw tools within an EA except for objects. Are you following The-Book to help facilitate your learning process. I personally stopped at Arrays and Custom Indicators and haven't fully grasped that section yet. This is like math class, you built on what you've already learned. Don't attempt custom indicator creations unless you're comfortable with all the concepts leading up to that point. Allot of trial and error will slow down your development process.

Speaking of math which I suck at. Does anyone know the equivalent of the following equation without using the devision sign? I want to avoid using devision sign as much as possible because it leads to a nasty error called Zero-Divide.

//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
//~~~~~~~~~~Lot_Size-Function
double Lot_Size(){Ans=0;
//~~~~~~~~~~
Sl_Price=NormalizeDouble(Sl_Price,4);
Median_Price=NormalizeDouble(Median_Price,4);
//~~~~~~~~~~
    Ans=(AccountEquity()*(Risk_Percentage/100))/
        (MathAbs(Sl_Price-Median_Price)*100000) /Order_Max;
    if(Ans<Broker_Min){Ans=Broker_Min;}
    if(Ans>Broker_Max){Ans=Broker_Max;}
//~~~~~~~~~~
return(Ans);}
//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
//~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

For example, I can replace (Risk_Percentage/100) with (Risk_Percentage*0.01). So how about the rest of it?

actually not that bad results, trough better exit conditions there might be room. Also i suggest adding an additional filter and allowing buy's and sell's simultaneously. I might give that strategy a look as well, even i don't like the macd ;)

Yeah, next stop MAE-MFE. I'll use the exit as is, then add random exits @ hour, 4-hour and day and see how the mae-mfe looks like. Which filter would you recommend. Time or Price-Action? All optimization will be done on this 3 month period and then we'll have a walk forward test, if it does well then we'll have a walk backward test.

Why not trade your system for some time. Generate enough money and pay for a tutor? This is what I was afraid of. The constant barrage of self-advertisement. You're still trying different things with your system. It didn't work good when you ran it through the night so you changed it .... then it started being flat to downward and you changed it again ... then it started losing and you ramped the lots.

Whats the incentive for someone to teach you to code for free. Allot of people come in here with Logs of Excel results showing 90% win-rate and even they cannot get a programmer to work for free.

You can also pay a programmer to code your EA for ya but you're afraid they're going to know your precious. You're in a real ketch 22 my friend. Give yourself one year from now and you too may be able to program complete EA's with all the stuff recommended here.

In the main-time, I'll take my little experiment out of your thread. Good Luck.