How I can know the backtest is not faked? It would be VERY easy to code it so it will use e.g. 30 minutes later data for calculating indicators than what we really are at the moment in the test time. So it would see a little bit into the future.

As I've said multiple times; EMA alone is not enough for making profit more than loss (without very good luck) so there is something suspicious with this "butter bot". I've ran hundreds of test runs in my own software with different kind of EMA strategies with multiple different EMAs at same time or not and it never did more profit than what it would've done by just holding the coins. When using few other indicators combined, it started making profit, but configuring takes time.

This post is not just to advertise my own bot. There is something suspicious with butter bot.