If so, please DM or post here if you feel comfortable. I do not want to talk about strategies, just about things like what’s your per trade expectancy, how many trades does your algo generate in a month, what kind of returns are you seeing, etc. I’ve recently had a breakthrough and my algo is performing very well in tests, almost too good to be true, and it has me immensely skeptical. I’ve spent 2 days poking holes in it and haven’t been able to uncover an issue yet so I’m trying to see if there is evidence of others who have achieved similarly outsized results.
For others who don’t fall into the target audience for this post, I respectfully ask that you refrain from asking/DMing for any secret sauce or piling on saying outsized results aren’t possible. In return, I will divulge that my strategy was developed using a machine learning classifier with only 1min candle and volume data as the core input, with various features then built off of this data. It took a lot of persistence and trial and error to see even some middling results, then 1-2 major breakthroughs (plus a big fix that was causing poor results even for the final model) to get to this point. It was very much a zero to one moment.
Submitted September 26, 2020 at 09:30AM by avabisque