Quick update on my efforts to automate the backtesting process (Patreon)

Here a quick status update on my efforts to automate the whole backtesting process to find the best algo strategy.

In my previous post I wrote that I was busy programming a complete backtesting engine that would try to fully automate the backtesting process. I am proud to tell you that, at the moment, my version 1.0 of this program is finished. 

After a couple of weeks this bunch of Python code does the following things for me so that I can concentrate again on finding and reverse enfineering other developers strategies. And also write my own strategies as well :-)

• I does a series of backtest over all the timeframes I usually tested manually (but can extend this in the future with other timeframes as well)
• It collects all the backtest data from the export json files and adds them to a SQLite database for further analysis (databases are created if they do not exist yet btw.)
• Then it calculates the score based on the scoring tables (more information about this concept in my post about creating your own strategy scoring spreadsheet)
• The best score is then automatically selected and added to the strategy league in another table
• The program also has the option to backtest and score the strategy after it has been optimized with Hyperopt
• It also makes profit plots of all the best scoring strategies so that I can compare the improvements/differences
• All important output goes directly to my screen (see picture above) and can also be used in video's and blog posts
• Finally it collects all data and stores it in a separate directory so that I can upload it to you as my patrons so that you can see these outputs and can compare these with your own findings

As said This whole proces significantly shortens the time and improves my whole work process. Each backtest is 100% similar to previous and future tests and the only difference is the strategy itself.

I expect to refactor and improve this program in the future and have more ideas for further improvements. But for now I am going to retest all the earlier tested strategies. I will add a zip file with the results of one test to this post so that you can see how the output will look.

In the mean time I'll keep you informed about this and other ideas I have to find the best algo strategy for Crypto.

Thank you for your support and your patience.

And I will see you soon!