Correlation coefficient of nearly 1 in a regression is not great. If the s&p returns swing up, and correlation not there, the return won’t be good. You really need it to be >1. I’m not an analyst nor an investor, just studied finance recently and find the topic interesting.