"While the 2003 book Moneyball popularized sophisticated statistics called sabermetrics in baseball, hockey analytics has lagged behind. In a National Hockey League game, each team scores fewer than three goals on average over three 20-minute periods. With around 23 players on each team and six men normally on the ice at any one time, it’s hard to tease apart individual contributions to scores. Figuring out the best lineup for each situation depends on having those kinds of data.
Compared with baseball, “hockey is just inherently more difficult to analyze. It’s very fast, very continuous, not as discrete, and there’s just not as many scoring events in it,” says Shane Jensen, coauthor of the recent review and a statistician at the University of Pennsylvania’s Wharton School."