Boolean

The close-to-optimal "go against the trend" ("buddha") trading algorithm I developed for my bot works in almost all instances. But "almost all" is not good enough. Some off days during Christmas I fiddled with some old algorithms I've tried out before. During a break when I went jogging I came to the realization that there are two things that I wish to address:

• I wish to take a much higher risk during normal idle/scalping periods, in order to get much higher gains; and
• I want to follow the trend during the occasional whale push. For two reasons:
• reduced risk, because opposing a strong trend risks getting you rekd, as I noticed all too well; and
• higher chance of profit, since my "buddha" algorithm dislikes trends and only really enjoys rebounds.

So I'll try and build something that switches between two modes: scalping and follow. When and how to switch will be the hard part.

You might say "why not run follow mode all the time?," and the simple answer to that is that you have to be extremely aggressive when following a strong trend. Just look at this 11% push to the upside:

From where you can tell this is going up-up-up, you have a few minutes to react, and you can't really put a low limit order and hope to get filled. You need to market order or hunt aggressively (which btw is close to impossible due to BitMEX getting overloaded). Aggressive trend following does not work well together with short-term, as you'll be buying high and selling low.

Now I'll do some manual chart analysis to try and figure out a repeating pattern, which tells me when to switch. Then implement and tweak. Then test-run in follow mode only. Then tweak them both together. Then launch. Then bug-fix. Peuh...