Robot Power EA

hey , actually i started forward test it last week. the results are good till now.

i also back test it since 1.06.06 - 18.10.06 and the result are good also (100% back profit in 4 months)

but when i tryed backtest it from 1.01.06 it lost all the money till 6.04.06 (4 months)

we know that the market isnt same like 1 year ago and that the problem with the experts ... we always tune it to now days , but the market changing all the time , year after year .. almost all the ea's u would found , if they r proffitble for a year on backtest , try to backtest it for 2 years , and u'll see whats happend ..

sorry for my english , here are the results of my forward test , i erase the trades that didnt made by robo, i use 4h chart

For some reason my model quality sux.

I can get great results and time of the past 2 years. But thats why I want others to test it too.

That one test is scary. We never know when the market will start doing that again. We can make adjustment to an ea for the past but we cann't adjust for future since we aren't fortune tellers. We gotta find settings that survive any kind of market conditions over at least a 2 year peroid.

Can you run that same bad peroid on 1 hr time frame same settings, to see what it does?

Last week was good for any trend folowing ea so last weeks forward test isn't really good judge of an ea.

what broker too?

thanks

Dave

ok I deleted all my history and reloaded it. Now its good.90%

I get the same now.

Problem with this ea is it always keeps a trade in. Soon as it closes a trade it opens a trade.

We need a coder to look at this ea and do something about that. That don't seem right.

Or it needs another indicacor filter. Like trend continuation, or I_trend. Something to trigger a trade and not kust keep one in at all times.

I think the problem is here in this code. Seems alot is not included in the entry calculation.

Any ideas?

void calculateIndicators()

{

bull = iBullsPower(NULL, 0, BullBearPeriod, PRICE_CLOSE, 1);

bear = iBearsPower(NULL, 0, BullBearPeriod, PRICE_CLOSE, 1);

Comment("bull+bear= ", bull + bear);

//sarCurrent = iSAR(NULL,0,0.02,0.2,0); // Parabolic Sar Current

//sarPrevious = iSAR(NULL,0,0.02,0.2,1); // Parabolic Sar Previuos

//momCurrent = iMomentum(NULL,0,14,PRICE_OPEN,0); // Momentum Current

// momPrevious = iMomentum(NULL,0,14,PRICE_OPEN,1); // Momentum Previous

b = 1 * Point + iATR(NULL, 0, 5, 1)*1.5;

s = 1 * Point + iATR(NULL, 0, 5, 1)*1.5;

// Check for BUY, SELL, and CLOSE signal

// isBuying = (sarCurrent sarCurrent && momCurrent < 100 &&

// macdHistCurrent < macdSignalCurrent && stochHistCurrent < 35);

// isSelling = (sarCurrent >= Bid && sarPrevious 100 &&

// macdHistCurrent > macdSignalCurrent && stochHistCurrent > 60);

isBuying = (bull+bear > 0);

isSelling = (bull+bear < 0);

isClosing = false;

}

when I try to take the // out from infront of the functions I get an error that those variables not defined.

Robot Power

I have been working on this expert for quite awhile trying to make it profitable,on this version added time filter,money management, and pivots avg sys, finished last weekend and opened a 50k demo with North Finance and heres the results, personally think I just happened to start it on a good week for this Expert but not sure. Seems best on eur/usd H4 timeframe for NORTH FINANCE set time to trade from 8 to 17.Renamed it Sixper because of six period pivots.On the demo results I had sixper on eur/usd only but then added it to the rest of the pairs except aud/usd and gbp/usd. Needs TTM Stoch in the experts/indicators and of course Sixper in experts file.

mrtools

cool thanks mrtools.

Is this thing suppost to keep a position in at all times? shouldn't it want to wait for a signal? because durring consilidation a bear signal can change to bull signal and visa-sersa.

and I keep getting this not enough bars error even though I have bars in chart set to max and max bars in history.

any ideas?

thanks

Dave

ok i fixed the always in thing.

Robot

Very good Dave been trying to figure out how to slow this thing down some,adding bulls and bears to the main buy and sell very easy fix.May make this a very good expert.

mrtools

Back testing takes forever and a day though.

edit. with the right time frame and indicaters set to the right settings I think it could ba good.

for right now I am testing with sar and macD taken out. Tests slightly improved on speed. Plus I don't like macD or sar. MacD gives too many false signals and I don't really understand sar much.

I think there's enough indicaters in it with out those 2. Just need good indicater settings.