Do I need to write my own "Strategy Tester" for MT4 (build 208) !!!?
I have also been experiencing problems with strategy tester. I've manually downloaded year 2003 to 2007's 5 minutes data from a trusted place and tested one of my EAs with the data. it worked just fine. I made sure by visually checking it trade by trade. this was in build201.
then I upgraded to build202, the EA fell apart. it still generated trades. but it simply wouldn't go before December 2005. that is, it wouldn't produce any trades for the period between 2003 and 2005.
this problem continues in build 203 to build 208. but as soon as I switch back to build201. everything works to a T.
this is just one of the worst problems with the latest builds of MT4. I've a few more.
then I upgraded to build202, the EA fell apart. it still generated trades. but it simply wouldn't go before December 2005. that is, it wouldn't produce any trades for the period between 2003 and 2005.
this problem continues in build 203 to build 208. but as soon as I switch back to build201. everything works to a T.
this is just one of the worst problems with the latest builds of MT4. I've a few more.
More info. Looks like a new bug for two reasons below (will report):
1. In my above historical test the ratio of M1 data with vol of 6 (a bug fixed in 208) is 7% (M1: 968870 bars, 74959 bars with 6 vol. Ratio=0.0774). I really don't think that could have affected my EA tests running on H4 that much.
2. Also this EA is running live on demo for 3.7 days and it has traded 7 times, so I really doubt that it will trade only 418 in my historical test for 3 years.
1. In my above historical test the ratio of M1 data with vol of 6 (a bug fixed in 208) is 7% (M1: 968870 bars, 74959 bars with 6 vol. Ratio=0.0774). I really don't think that could have affected my EA tests running on H4 that much.
2. Also this EA is running live on demo for 3.7 days and it has traded 7 times, so I really doubt that it will trade only 418 in my historical test for 3 years.
==========Following is the answer I got==========
Hello Shahin,
Please refer to the articles at https://www.mql5.com/en/articles/mt4 You can also refer to the community at https://www.mql5.com/forum/en for the help.
Best regards, Tatyana
MetaQuotes Software Corp.
https://www.metaquotes.net/
"MQL4: automated forex trading"
Hello Shahin,
Please refer to the articles at https://www.mql5.com/en/articles/mt4 You can also refer to the community at https://www.mql5.com/forum/en for the help.
Best regards, Tatyana
MetaQuotes Software Corp.
https://www.metaquotes.net/
"MQL4: automated forex trading"
I received your letter. Please provide the source of your EA and two tester's reports from buid 207 and build 208. Also please inform in details how you downloaded the historical rates "5 minutes data from a trusted place and tested one of my EAs with the data."
I download and imported all data from Alpari Russia. One minute data is used for H4 to build the history for tester (as it said --using M1--). All data starts and ends at the same time (extra is deleted), for a length of 3 years.
I am hesitant to send you my source.
Build 207 & 208 tester report: (in email--you can add here if you like)
I am hesitant to send you my source.
Build 207 & 208 tester report: (in email--you can add here if you like)
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
Could you help with the strategy tester? Or make that part Open Source so experienced programmers like myself can work on it and fix the bugs?
I mean look at the results below (same data = data has not changed at all), except maybe a little on swaps.
Modeling quality (on both builds 207 & 208) =88.15% every tick H4 (using M1), Bars in test=4853, Ticks modeled=800 ticks less in build 208
Build 207: Profit factor=18.16, Profit trades%=96.62, Total trades=5533, Total net profit=685613
Build 208: Profit factor= 0.88, Profit Trades%= 66.27, Total trades= 418, Total net profit= -3846
I read the news and there were no talks about such discrepancies, what is going on? Last time "BIG difference in Strategy Tester from Build 198 to 200 (bigger in M1)" this happened metaquotes said: "More accurate modeling" -- Please say more.