Stefano Soccorsi pointed out that the sequential moments equation on page 19 is wrong. It should be:
1t∑i=1txim.