AI - BTC (1) After discussing gold, let's continue to take a look at some of the research AI has done on BTC in the past period .. The first one is certainly the same as the golden idea. If multi-agent trading can run smoothly in gold, can BTC run smoothly? After all, there are more various data on BTC .. I really searched for a lot of data on this board I almost used most of the things in MMT's $499 API for a month Create this board . It is also a multi-agent approach, including ICT/Price Action/Liquidity/Scalping, and so on . The final result is a bit difficult to evaluate .. Although after a month, PnL can barely run straight But the process was very bumpy, basically in a state of open and closed If you want to make a big profit, you can also make a big loss . When it's good, the total PnL breaks 20k (analog disk), and when it's bad, it can also be pulled back from a PNL of 20k+to over 6k now .. Overall, BTC is still in an unusable state at present ... It still needs to be reset and redone We have paused running now, and we will plan a new set for the next 1-2 weeks .. But among them, if you look at Figure 1, you will find that the only agents that can run well in both winning rate and PF are actually two scalping agents At that time, there was an idea that if everyone's agent's frequency increased from three times a day in Asian, European, and American markets to once every 4 hours or even every 2 hours Will the actual results improve? So I researched another BTC scalp stripping system .. After running intensively for a few days, it doesn't seem like much .. (Figure 2) Burning tokens every 2 hours is a bit too harsh Also paused for now So currently, BTC's trading agent is also in a suspended state In addition, it seems that MMT's API payment end is broken these days and cannot be renewed After recovery, we will study other paths . Next, let's talk about other ideas that BTC has done