Tickstory

I could find quite big spike at EURJPY 2007-04-04 13:40:04 to 13:43:21 and it makes EA testing weird.

Even if this similar thing happened at that time, I can't rely optimization with this data as it increases my earning incredibly and I can't compare optimized value between them.

Can I modify or delete these data?
Is there any useful link or info?
Maybe exporting to a readable file and convert to mt4 format would work?

Hi Redred,

Unfortunately there's no easy way to modify the original Dukascopy data. The first thing you may try is to re-download the corresponding data files (*.bi5) by firstly deleting the corresponding dates/times from the repository. The directory structure takes the form:

<instrument>/<year>/<zero-based month>/<day>/<hour>.bi5

If the newly downloaded data still has the problem, then you can try a bit of a fiddly workaround by writing a zero-length file over the top of the offending hours.

Hope this helps.

tickstory wrote:Hi Redred,

Unfortunately there's no easy way to modify the original Dukascopy data. The first thing you may try is to re-download the corresponding data files (*.bi5) by firstly deleting the corresponding dates/times from the repository. The directory structure takes the form:

<instrument>/<year>/<zero-based month>/<day>/<hour>.bi5

If the newly downloaded data still has the problem, then you can try a bit of a fiddly workaround by writing a zero-length file over the top of the offending hours.

Hope this helps.

I made that b15 file zero-length and exported to mt4 again but price didn't change.
Local tickdb file is still zero-length. I don't know why this behaviour happens.

At last, I modified EA to skip any action at certain time period to workaround this problem.
Now I can see smooth profit graph.

Thanks anyway.