When a trader browses the historical data of Lido Machinery (the virtual company of this article), it will be more convenient to find out at which stage the stock has a P/E ratio between 15 and 25. Since this data is generally reflected in the company's financial reports, it is necessary to look at the financial reports for the past quarter, half year and full year for several years. This work is very laborious and time consuming. After finding the eligible data in a large number of reports, it is necessary to link the data to the price fluctuations of the stock, which is a very time-consuming and laborious task.
Another issue that traders need to address is how much the sample size for historical data testing should be appropriate.
There are two extremes in the choice of sample size: one is to select the largest number of samples as much as possible; the other is to select only one or two samples. By describing the pros and cons of the two extreme conditions, traders can better grasp the appropriate workload.
As many samples as possible will make the trading strategy fully tested by historical data, but its workload is very difficult for individuals. Think about it, if independent traders face financial reports from more than 3,000 listed companies for several years, this will be a huge workload. Even if there is perseverance to collect these reports and test them, the market and many stocks may have changed a lot.
If the sample is only selected one or two, although it will be very easy to pass the historical data test, it is only a great chance that the test results can prove the universality and reliability of the strategy.
(to be continued)
For more articles, please browse "English version index"
没有评论:
发表评论