modeling confussion: use 1 minute Strategy Testor for max EA test accuracy yields 25% MODELING QUALITY
I was just reading this article 'Tester in the Terminal MetaTrader 4: It Should Be Known' because I too was wondering why I was only achieving 25%.
According to this article that is the best you can get with M1 data.
I was just reading this article 'Tester in the Terminal MetaTrader 4: It Should Be Known' because I too was wondering why I was only achieving 25%.
According to this article that is the best you can get with M1 data.
Thanks so much for your reply
From what we see in the article, I presume that the proper period for the strategy tester would be the highest timeframe being tested. As a NOTE: I have proven that the same trade results occurred on the M1 timeframe as the H4.
So my question now for the moderators, phy etc. then is:
When an EA scans multiple periods, should the strategy tester be set to the largest time frame scanned?
Now that I have optimized my code across individual time frames, I want to be able to now test the AE as it scans through the time frames ie M30 H1 H4 D1, Should I use time frame D1 or time frame M30?
Regards,
John
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
After much trails trying to get reliable historical data...
I have migrated to a new Alpari leaving at build 223.
In an article it suggested that once history data has been acquired it is far more accurate to use int value rather than Period() function. This was of no consequence because my EA looks across many timeframes. The EA is tailored to larger timeframes specifically H4 but was not yielding the results expected. Trade attributes for determining have been intermittent.
When looking at the report I see that the modeling quality is rated at 25%.
When I run the Strategy Tester at H4 I get the same trade results somewhat proving that running the tester at 1M is valid; however, I do now get modeling quality of 90%.
What's up with that?
Many thanks to reasoning
John