Understanding spurious correlation:
Interesting and funny, thanks for your post 
Harrogath,
Thank you. For sure we should take the length of the series that could be very short to be significant. It also shows the need for using a confidence interval to appreciate the likelihood of such occurrence.
Confidence intervals do not allow one to make probabilistic statements about a hypothesis or the true parameter value. The only probabilistic statement one can make is about the long-run performance of the methodology. For example, a 95% confidence interval for true mean stock return is 3% to 9%, however, this does not mean there is a 95% chance (or likelihood) that the true mean return is between 3% and 9%.