CryptoPainter

Hard work pays off, just like water drops wearing down an iron rod into a needle, and stones pierced by a determined person! Friends who have worked with genetic algorithms should be able to see that my system has finally run successfully. Two point arrays represent all surviving strategies in the current market environment converging into two dominant species, and they are still continuously iterating and evolving... And the environment they survive in is the current BTCUSDT - 1min market! This is not backtesting, but real-time trading! In other words, as long as you provide high randomness and a large number of species, even in a market with an extremely low signal-to-noise ratio, strategies adapted to the current market can evolve through long-term survival of the fittest. Of course, this does not mean the strategies have succeeded, but the process has run through. After all, these dominant species have only survived for 3 hours. After a sleep, the dominant species will probably have changed... What I am waiting for is the cockroach among these strategies, that is, a strategy that can survive long-term. Although theoretically it does not exist, it doesn't hurt to try! Because the entire system's evolution is not done through backtesting, theoretically the strategies have minimal overfitting. We don't need to select the optimal strategy through this system; we only need to continuously let the live trading execution refer to the signals given by the current dominant strategies! Once a dominant species declines, decisively abandon it, so the entire system's trading logic becomes completely dynamic. It's really fun!