why the the detected sequence add a interval*2 margin when plot the result?

[hellopari]

Hello, when I read the code and the finally anomaly detection results figure of your paper, I find the axs.plot( j * interval * 2 ),i cant understand why. If i remove this margin, the detected anomaly will cant cover the true anomaly. So can you explain why? Thanks a lot!