Disliked{quote} Hi sevan62, Talking about data accuracy, I'd like to get you opinion about this: I've tested an EA with pepperstone data. The data quality was 64% in the results report. In order to get this data I've loaded all data possible provided by pepperstone in all timeframes. Then to get more data ticks I've loaded data from metaquotes using the History Center. Then I run the test again and got 90% data quality. I ask you if the MT4 loads the data setting up the GMT offset of that data with of the broker. Is this result reliable? If you have some...Ignored
We are in a Renko trading thread now but i hope you are posting about usual TF (not Renko) bactesting coz Renko backtesting is another big story.
- The fist is the most important thing is that almost all MT4 brokers cuts off M1 downloading history up to 2 months. It is a period on which you may have the maximum avalable native MT4 backtesting quality of 90%. You have to switch to M1 TF and press PgUp or Home button and wail till all M1 data would be downloaded. Another way is to use scripl forcedloadhistory which do the same work but without pressing a button. If you want to have a longer M1 period (more than 2 months) you have to open MT4, M1 chart and stay connected to brokers server as long as you need. You may also connect from time to time but with pauses not more than 2 weeks.
- In case you load data using History Center you should know that this data has nothing to do with your brokers data. Histoy Center data is MQ (MetaQuotes) data and it is the worst data i have ever seen. It is not reliable while backtesting using MQ History Center. Yes while loading from MQ History Center MT4 would adjust time offset according your broker GMT offset.
- The last thing which have a great effect on backtesting is EA principles. The most accurate and the fastest backtesting may be obtained if your EA makes all actions (open/close and trail postition) on close bars only. And you dont need to to care of 90% modeling quality in this case. You should have only the data for TF you are backtesting and you may use the fastest model "open prices only". The highest data quality requirements when you are testing on M1 TF and EA may open/close or trail position on the same bar.
BR, alex