I created a database from 1 minute data, and am learning AmiBroker by using daily bars. The charts look good, and I am trying to understand more about how daily bars are constructed from 1 minute bars, and how the fact that daily bars have different start/end date for different symbols could affect testing.
When not using pad/align feature, is it ok for "# bars" to be 8-9 in a week? I don't mind having that many, since it's pretty clear those are not full-day bars. But I'm wondering if some other areas of the backtesting process would be affected. And, if it's not ok (i.e. it should be 5), what should I check or do with my data? Maybe I need to import it in a different way? (all I imported was 1 minute CSV files).
When testing individual symbols the # of bars is 5 for both, but when testing multiple symbols as part of a portfolio the # of bars changes to 8 or 9 (see picture below). I think I understand why (there are 8 or 9 intervals that result from the combination of start/end minutes of the component bars of the two symbols), but it's just a guess, so I'd appreciate someone with experience bringing clarify to this matter.
I also tried enabling pad/align, and set EURUSD as the reference, thinking I have the most complete data for that. As seen in the picture below, when using EURUSD as reference the # of bars is shown as 5 (rather than 8 or 9 when not having any reference).
If using a reference is the best approach, I am thinking of creating a CSV file that covers all possible minutes in the historical data base, and use that as reference symbol. That way I will avoid cases where my gaps in EURUSD data affect other symbols.
Below are 2 pictures that I hope clarify what I'm talking about. Thanks in advance for any comments or suggestions.