In the last article, there was the mentioning of splitting historical data into 3 parts. One-third needs to be reserved for out of sample testing whereas two-third would be used for in-sample testing.
Out-Sample In-Sample Forward Performance
\ /|\ /| /
As we can see above, the distribution of historical data. Here comes the role of correlation. It refers to the performance and overall trends. Correlation metrics can be used while evaluation trading strategy performance reports created in the testing period. The better or stronger the correlation between the two, the higher the chances of success of the performance of the system in forward performance testing or in real-time trading.
Once a trading system has been developed with the use of In-sample trading data, it then is ready to be applied to out of sample data. Providing traders an option to evaluate and compare the results of the performance between In-sample and out-sample data.
If a strong connection is seen then the next step is to involve the addition type of out of sample testing also known as forward performance testing.