I have looked everywhere however I can't find one example of a worked auto-covariance solution using k lag. For example say I have a list of values: [7,10,15,22,19,28,20,17,33,29,37,35,40,32,44,47,38,31,28,22]. How do we compute the auto-covariance vector using lags of k=0,1,2 and 3?