And need help
I am going nuts.. LOL, I just cant get my strategy tester
to give me the right results.
Can someone please do a backtest. 1 year from today back
1 year using this ea. Please do several pairs if you can
on 4 hr and 1 hr. Please!!
Any other pair you like??
If someone can help..thank you so much!
forget it !
Please forget this strategy tester.
you can have a financial market is not free.
Please make them take a demo account for 10 currencies and test 30M, H1 and H4. a month they will see whether the ea and win brings works or not!
What time frame do you prefer? Or do you use different time frame
based upon which pair?
Yes agreed...I think I am taking strategy tester with less
than a grain of salt.
Originally Posted by viennatrader
First I also thought back testing is totally unreliable till I witnessed one trader on tsd forum making great weekly optimizations that do work in forward tests. He told me, one has to collect tick data for that. Just 1min data or Historical MQ data isn't going to be back/forward test reliable, you need tick data continuously collected for few months.
Then you get somewhat reliable back test results. Also one has to compare forward tests versus optimization pattern (data duration/variable selection). There are certain optimization patterns that show even greater results on forward tests. So this is more of an art.
Of cause this applies to quantitative and sophisticated rule trading systems like nero nets that rely on weekly optimizations.
Simple rule or data action systems need forward testing
Last edited by Metatrader7; 06-22-2009 at 18:57.
I don't get a smiley face when I installed this, it is in the EA folder.
Is there any file that I am missing?
How could backtesting be unreliable. How else can you see how your system would perform in past trading, if something has done well in the past then there is a greater chance it will do well in the future, only common sense.
I optimize all my robots on a monthly basis and they continue to be profitable for me, how does one know what imputs to use unless you optimize the robot, the same imputs do not work for all pairs and time frames, they are all different and as such need continued optimizing.
Since the market semtiment changed quite quick, the only way to keep a head of the pack is to optmize on a regular basis, like i said above i do it every month MAX!
Originally Posted by Metatrader7
I am doing forward testing on this EA on GBP/USD time frame of H4 in my VPS. I will publish the result from time to time at Forex Review Empire. Check it out for more data. Thanks
I tried forward testing for a week and publish my result in my blog Forex Review Empire with positive earning but only one trade for the whole week. Anyone getting the same result? Thanks
Both back/forward testing are needed
Originally Posted by wannaberich
Agree, truth is somewhere in between,
To take real advantage of back tests one has to forward test and compare both methods, then you can tell what back test does and what it misses. Then you know what variables can be optimized in back test and which aren't. It does vary per system. It requires lots of tweaking experience and skill to do it right. Not all manual traders are adept tweaking wares. Not all aspiring programmers have enough trading sense either.
So blind reliance on back tests without considering forward test results does no good either, cause results can be back fitted or overoptimized. One has to keep in mind both ends of the story.
That adds complexity, not everybody can handle that. Pure forward testers are closer to real trading though even lacking back testing skill.
P.S. Who said trading is difficult? It is simply complex. HaHa
Tags for this Thread