Backtesting/Optimization - page 10
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
I had the same problem but seems I fixed it. But no any unique decision for this because the problems may be different.
For me for example was the following decision. I filled 9999999999999 for history and for chart as well. Besides I opened offline mode (menue-> File) and I saw that I have two files as M1 with different size/date of data and so on. So I had to import the good files once again from file forder to the same file forder and it works well now.
Metatrader is not doing 90% automatically for you even if you did everything in right way. So it is necessary to set "Use date" option and play with the "From" and "To" dates to get 90%.
I tried both your methods and I still got only about 56% quality. Personally, I think everybody should tell their brokers that they will quit trading for a week if they do not get a response from MetaOrg to promise to fix this lousy backtester within a month or two. Then this frustration would be fixed once and for all. Unfortunately, forex traders will not stick together on this issue. This messed up backtester has been a joke for over 2 1/2 years now. I wish someone would create one seperate from Metatrader that can test anybodies ea with 98 - 100% accuracy. I guess there are not enough serious traders out there. Most must like to keep losing money on the forex because of lack of proper and accurate ea backtesting - Very sad!
I tried both your methods and I still got only about 56% quality. Personally, I think everybody should tell their brokers that they will quit trading for a week if they do not get a response from MetaOrg to promise to fix this lousy backtester within a month or two. Then this frustration would be fixed once and for all. Unfortunately, forex traders will not stick together on this issue. This messed up backtester has been a joke for over 2 1/2 years now. I wish someone would create one seperate from Metatrader that can test anybodies ea with 98 - 100% accuracy. I guess there are not enough serious traders out there. Most must like to keep losing money on the forex because of lack of proper and accurate ea backtesting - Very sad!
I know that it is not easy.
I spent more than 2 weeks to do it: I install/deinstall Metatraders and so on.
It was not easy.
Try this once again http://www.metatrader.info/node/67
Besides it may be good to have one copy of Metatrader installed especially for backtesting. Besides when you are doing preparation to have 90% it may be good if you disconnect from the server (to avoid any mixing the data which you are importing and broker downloading in the same time). Install fresh copy of MetaTrader, open an account, disconnect from the server immediately after that, delete downloaded history files from Your_Broker_Demo folder (just some of because there are some files there which can not be deteted), download your imported data and follow this article http://www.metatrader.info/node/67
I did it many times untill I got 90%.
I had a lengthy discussion with my older brother about the 90% modeling quality. He said he had the same problem with a software package that he used in trading stocks. After a couple of months the originators of the software fixed their backtester where it was 98% accurate to live trading during the same time period. Why the "H" are we settling for garbage then?? 90% stinks also.
Again, what are we trying to do, play at making ea's or are we serious about troubleshooting ea's that we can guarantee to an extremely high degree that they will work as tested on the backtester and we can then make actual money at this game? For the life of me, this totally blows my mind that people are not protesting this worthless backtester. Is this the norm that people just want to lose money on ea's in actual trading?? I must be talking to the wall, otherwise people would be protesting this joke of a backtester to their brokers or to Metatrader.
Dave i dont think alot of people on here understand the importance of 90% + quality data for back testing
The number of posts i have seen with EA's that have 40 or 50% quality.
And i agree i dont think a lot of people are that serious, because you would have to have one hell of a good EA and the programming of such is probably beyond MT4 capabilities.
please note i am not an expert programmer or anything like, this is just my opinion from what i have read and seen on here
I had a lengthy discussion with my older brother about the 90% modeling quality. He said he had the same problem with a software package that he used in trading stocks. After a couple of months the originators of the software fixed their backtester where it was 98% accurate to live trading during the same time period. Why the "H" are we settling for garbage then?? 90% stinks also.
Again, what are we trying to do, play at making ea's or are we serious about troubleshooting ea's that we can guarantee to an extremely high degree that they will work as tested on the backtester and we can then make actual money at this game? For the life of me, this totally blows my mind that people are not protesting this worthless backtester. Is this the norm that people just want to lose money on ea's in actual trading?? I must be talking to the wall, otherwise people would be protesting this joke of a backtester to their brokers or to Metatrader.
Hi,
Backtest is nothing!
I'm running two EAs and had only 2 orders in 7 days. In the backtest with MQ 90% I had 11 orders so backtest means nothing. That's my opinion.
Have a nice day
I guess there are not enough serious traders out there. Most must like to keep losing money on the forex because of lack of proper and accurate ea backtesting - Very sad!
I could not agree with you more.
I have seen how powerful good backtesting and optimization can be in system development. I have seen optimization used to guide the development of systems that simply wouldn't be possible without the optimizer's guidance. Here's a bold statement that I firmly believe to be true: A good backtester/optimizer can show you things in the data that you just can't see without it.
I've been watching a proof of this unfold slowly over months. I saw a system get developed on another platform in about 2 weeks, from start to finish. The system is hot. (And proprietary, and not for sale...) It was built with the help of an optimizer. Meanwhile, I've watched a number of people work together on several MT boards for months now, gradually building a system thats based on a similar 'original idea'. Despite the serious, concerted efforts of a number of participants, forward testing, coming up with new logic structures and filters, forward testing some more, the current best versions of the MT system can't touch the '2 week system' and I don't think it will any time soon. I'm pretty sure one of the first versions of the optimizer-built 2 week system was better than the current best MT version, after months of effort.
Basically, I think the MT world has no idea what it's missing. Without a working optimizer, a whole dimension of possibility simply doesn't exist. Those with extensive experience with an accurate optimizer will know exactly what I'm sayin'...
Meanwhile I keep reading MT people saying things like "No system has been proven to work consistently"; and just today I saw somebody saying that the "Holy Grail" is a "...'pot of gold' at the end of the rainbow. Thats right, it doesn't exist!!" As way wrong as these guys are, I can't blame them, cause they've never seen what's possible. They have no idea...and without a working optimizer, they aren't likely to get one any time soon...
- - -
I don't want to convey an idea that a working optimizer is the sole ingredient needed to finding a holy grail system. And, an optimizer may not be necessary to find one at all. But it certainly can be a major, major help.
Hi,
Backtest is nothing!
I'm running two EAs and had only 2 orders in 7 days. In the backtest with MQ 90% I had 11 orders so backtest means nothing. That's my opinion.
Have a nice dayIt's because backtesting is the testing EA on the past data. Forward testing is testing on the real data. But when you finish forward testing (forward testing day for example) your data is becaming a past one.
If we agree that EA (bases on some system) will perform in the same way as it performed 1, 2 or 3 years ago so it is ok. I mean that there should be some sequence in EA's working (because on the indicators and trading systems) so we may "interpolate" the past results on the real on (on the future i mean) just to understand how it may be and to check the trading system or algorithm. Each EA/tradibng system is having some algorithm so backtesting can prove it and we may see how this algorithm is working.
But it is necessary to have 90% to do this.
It's because backtesting is the testing EA on the past data. Forward testing is testing on the real data. But when you finish forward testing (forward testing day for example) your data is becaming a past one.
If we agree that EA (bases on some system) will perform in the same way as it performed 1, 2 or 3 years ago so it is ok. I mean that there should be some sequence in EA's working (because on the indicators and trading systems) so we may "interpolate" the past results on the real on (on the future i mean) just to understand how it may be and to check the trading system or algorithm. Each EA/tradibng system is having some algorithm so backtesting can prove it and we may see how this algorithm is working.
But it is necessary to have 90% to do this.Ok, you are right but tell me one thing. I'm not a programmer... I have an EA that give me only MQ 24% so where can be the problem? Every other EA has 90%.
Thanks in advance
EDITED:
Only when I put the EA to work with TF: M1
It is apparent that for serious traders, the backtester is a complete failure on the part of Metatrader. Without backtesting over specific market conditions, one never quite knows how an EA will react in the future - That is why we need a backtester. 90% (If you can ever achieve it using this strategy tester) is worthless. It is a crying shame that this system was developed in Russia. Evidently, they do not care enough about their reputation to fix the system - That says volumes about their integrity. It makes me wonder what kind of people are behind the Metatrade organization??
A system that can provide 98 - 100% accuracy as compared to actual trading during the same period of time would eliminate thousands of hours of time consuming forward testing. I just wish we could apply pressure and demand that they fix the program or we would quit trading using the Metatrade platform - That seems to be the only way I can think of to get them off their behind side and produce a quality backtester product.