I will be using https://api.darwinex.com and Python and upload here my code in order to do realistic backtesting for Darwins. Although Darwinex already has a tool for Darwins backtesting, my believe is that it suffers from look-ahead bias. I will try to prove it with these snippets of code

https://drive.google.com/open?id=1jrqvYkTx-pMBPtICPCLkUxPgdzudgPqx to download get-badges output (arount 400MB file, and more than 8 hours downloading)

This function uses pickle to store an object (save_this) in binay format. File name will be <file_prefix%Y%m%dT%H.%M%S.bin> The function returns the string of the filename

This function uses pickle to load the last of the files in the current path (given by os.getcwd()) with file_prefix within its file name. The function returns two things: the filename and the object, or -1,-1 in case of not finding files with file_prefix in the file name.

This function gets basic info for all darwins. Its parameter is to force download from Darwinex, or to let the function load previously downloaded data. The function returns a list with all the darwins in it.