Recently, I saw that the first season of the AI Cryptocurrency Trading competition organized by NOF1 has ended, and the results are quite interesting.

Six AI models, with a total of $60,000 in capital for real trading. At the end of the period, the account only had $43,000 left, collectively losing nearly 30%. To be honest, this result is a bit tragic.

However, there is a detail worth pondering - in the contract grid area, the AIs surprisingly achieved "team wipe survival". What does this mean? It means that although the overall loss is present, those few AIs using the grid strategy are still doing relatively well.

What could this indicate? AI, in terms of engaging in short-term operations that chase price increases and sell during price drops, may not be very effective. However, if it is allowed to implement a neutral, mechanized grid strategy, it might actually demonstrate systematic advantages.

After all, machines do not have emotional fluctuations and strictly follow rules. Grid strategies inherently profit from sideways market conditions, without the need to predict directions, which perfectly suits AI logic.

Claude finally took the championship, and how exactly he did it is quite intriguing.