In your book, Cybernetic Trading Strategies, Chapter 8 discusses using Pearson's correlation to improve classic intermarket relationships. One example given is using a simple correlation between T-Bonds and the S&P500 for the period from 4/21/82 to 7/26/96 (page 121.) However, in that period, as today, T-Bonds and S&P500 did not always trade on the same days. Some recent examples of T-Bonds trading while S&P500 did not are 4/6/07, 10/8/07 and 11/12/07. Pearson's correlation compares two samples of an equal number(length) of data. If the independent sample has 15 data but the dependent has only 14, the result will be erroneous. And it can take quite some time, depending on the length setting, for the result to get back on track and be correct. My question is, how do you deal with this problem of differing trading days?

Just like tradestation is will use the last 15 data points for both so for data 1 it's

Jan 3 to Jan 17

for data 2 it's

Jan 2 to Jan 17 , with one missing day.

Hello, great to read informative stuff which contains very enlightening information. I can see so many people visit this blog and contribute to the meaningful comments and exchange of thoughts and ideas that it facilitates for me..
Thanks & Regards
Cynthia Brown
Datarecoverysoftware successful resolution