{x} | Trading.live

To begin with, we summarize the lifecycle of a trading model, which is illustrated below in a normal distribution. In the graph, the Y-axis represents the profitability or probability of profitability of the trading model, while the X-axis represents the timeline. There are several crucial points on the graph that traders must clearly define.

Firstly, at point A, a trading method (or trading model) must exhibit a period of profitable performance before traders in the market discover it. Through the process of random back-testing, traders discover a certain trading method that proves to be profitable based on historical data, leading it to be located at point A in the lifecycle.
The second critical point is point B, which is based on the premise that all trading methods will have a point of highest profitability and lowest risk. As the market trader structure shifts, the trading method's performance gradually declines, leading to reduced returns and a worsening risk-reward ratio. It's essential to note that once traders enter the market, they face certain risks.
Finally, at point C, the trading method's risk-reward ratio has deteriorated to the point where it can no longer provide excess profits for traders. Even if the trading method can cover transaction costs, it cannot compensate for the risks that investors face. Hence, traders must avoid executing the trading method after point C.
After defining the key points in the life cycle of a trading model, I would like to share a common phenomenon that I have observed based on my experience. Novice traders tend to develop their trading models near point B in the life cycle, as this is when the historical data of the model is most abundant and its performance is continuously improving. For instance, a trader may discover a trading model, Model A, with excellent performance in the past three years, exhibiting a year-on-year increase with an annualized return of 31% and a maximum drawdown of only 6%. Based on its strong performance, the trader decides to test it with real money. However, after six months of actual trading, the annualized return rate of Model A drops to 17%, with the maximum drawdown increasing to 10% - a typical phenomenon of the model's profitability deteriorating and the risk relatively increasing after point B. After a year of actual trading, the annualized return rate of Model A further drops to 6.3%, with the maximum drawdown reaching 14%. At this point, the risk-reward ratio of the trading model becomes unreasonable and it should not be continued.
This is a common situation faced by novice traders when developing trading models. They tend to favor models with better back-tested performance and subjectively estimate a discount on expected future returns. Over time, they may find that the trading models they develop perform well in back-testing but poorly in actual trading, leading them to constantly develop new trading models. Unfortunately, the more complex the trading model, the narrower its life cycle curve (similar to the shape of a log-normal distribution), and its performance declines rapidly.
The lifecycle curve of a trading model typically shows variation speed from point B to point C, which is dependent on changes in trader behavior in the market. External factors such as financial policy environment,/>/>/>/>/>