Skip to main content

Everybody in the (Data) Pool

Before diving into the deep end of the hockey analytics pool, we wade through the fundamentals

by Bob Condor / @NHLSeattle_ / NHLSeattle.com

Before dipping even one skate into the hockey analytics pool, let's start with three guys who dove straight into the deep end. First up is Frederick Winslow Taylor, a mechanical engineer from Philadelphia who first applied analytics and other engineering principles to promote efficiency in manufacturing processes in the late 1800s. Taylor touted time management as an important factor to work output. He basically founded industrial engineering as a scholarly and professional pursuit. 

Second up is Henry Ford, who applied analytics to his automobile assembly line. He was most interested in measuring the speed of work. Ford's success inspired more businesses to analyze how to be both more efficient and productive-the practice of "analytics." Computers supersized the analytics movement because more data could be calculated and analyzed way quicker. By the late 1960s, computers were being used regularly to support business decisions. Business analytics were here to stay in the business world, leading to data analytics that can be simply defined as researching, discovering, formulating and interpreting patterns within data. 

The third guy in our analytics founders group is Billy Beane, credited by many statisticians as the key figure in popularizing sports analytics. As general manager of major league baseball's cash-poor Oakland A's, Beane researched "sabermetrics" (named for an association of then baseball statistics hobbyists) and decided the concepts provided a potential competitive advantage because, among other new-generation stats, sabermetrics valued on-base percentage significantly more than typical MLB touchpoints as batting average, home runs and runs batted in (RBI). 

Sabermetrics posited that teams with players sporting higher on-base percentage (hits plus walks divided by at-bats) are likely to score more runs. Beane's plan worked: The A's made the playoffs despite one of the lowest payrolls in the league. They went on a 20-game winning streak during the 2002 season, catching the nation's attention as well as best-selling author Michael Lewis, who was inspired to write the book "Moneyball" (2003). The term "moneyball" stuck as a sports analytics nickname and the concepts went far and wide in the sporting world: Soccer, basketball, hockey, football and more. The esteemed Massachusetts Institute of Technlogy (MIT) holds an annual sports analytics conference renown for its scholarly work and sports personality guest speakers. The movement gained even more fame when Brad Pitt starred as Billy Beane in the 2011 "Moneyball" movie.

Sports analytics bloomed in the NHL during the just-ended decade. In 2009 Chicago Blackhawks general manager Stan Bowman quietly hired an outside analytics firm to specifically help him find under-the-radar players at lower salaries to support a core group of promising young players such as Patrick Kane, Jonathan Toews, Duncan Keith and Brent Seabrook. The team started charting player performance more closely for puck possession with less emphasis on goals and assists. Simply put, puck possession is control and movement of the puck for as long as possible. The only time losing puck possession is a positive is when the player or team takes a shot.  

Puck possession is most obvious to fans during a man-advantage power play, when one team choreographs what looks like a game of keep-away but is actually a strategic way to get players into the best scoring position. If you watch the Washington Capitals on a power play, expect the Caps to be find a way to set up surefire Hall-of-Famer Alex Ovechkin in his "office" at the left faceoff circle. 

Keeping control is much harder to do during the standard five-on-five skaters action during a hockey game. Defensemen carrying the puck out of the defensive zone are often badgered or "forechecked" by the opposing forwards. In the center or "neutral" zone between the blue lines that mark the zones with nets, players attempt to break up passes or physically disrupt the players rushing up ice with puck.

In those zones with the nets (one team on offense, the other on defense), there are scrums in the corners behind the goal line, checking along the boards, sticks/legs/bodies stopping passes between teammates and more. Computer and video technology allows NHL teams to easily record puck possession for players both as individuals and as a line or defensive pairing. Hockey coaches at all levels preach puck possession, including such as regroup, flow and puck support. It is effectively all about creating high-quality scoring chances plus the obvious bonus factor of keeping the puck away from the opponents attempting to take their best shots on goal. 

The analytics outsourcing worked out well for Bowman and Chicago. The franchise won three Stanley Cups during the decade (2010, 2013, 2015) using "Corsi" ratings as an early-on advanced statistic to scout hidden-gem players who helped their teams control the puck even the scoresheet didn't show copious goals and assists. Other NHL teams with high puck possession numbers during the decade: Los Angeles (two Stanley Cups) and Boston (one Stanley Cup and three Cup Final appearances). 

Corsi measures shot attempt differential while at even-strength or five-on-five play. The stat's elements include shots on goal, missed shots on goal and blocked shot attempts towards the opposition's net minus the same shot attempts directed at your own team's net. Chicago star Patrick Kane was known for his puck possession during the Cup years. Former Seattle Thunderbirds favorite (now New York Islanders all-star forward) Mathew Barzal ranked highest in the during his NHL rookie season of 2017-18. 

Corsi was created by a Chicago financial analyst, Tim Barnes, who published his formula and findings under the pen name of Vic Ferrari. He listened to a radio interview with former Buffalo Sabres general manager Darcy Regier discussing his team's shot differential (basically average shots for per game minus the average shots against per game). Barnes went to work on better ways to measure shot differential. He wanted to name the stat after Regier but it didn't seem to have the right ring to it, same for then Sabres head coach Lindy Ruff. So Barnes scoured the Buffalo staff and picked out Jim Corsi's surname because he liked long-time NHL goaltending coach's mustache. 

Corsi has enjoyed a successful career mentoring any number of all-star goalies, from 2014 Hockey Hall of Fame inductee Dominik Hasek to current Columbus rising star Joonas Korpisalo. His namesake stat has been reinvented, reworked and renewed into a wide swatch of advanced analytics such as expected goals percentage, zone entries and high-danger scoring chances. All of which we will examine in future dips in the deep hockey analytics pool.

Congrats on finishing your first data-swim lesson. 

View More

The NHL has updated its Privacy Policy effective January 16, 2020. We encourage you to review it carefully.

The NHL uses cookies, web beacons, and other similar technologies. By using NHL websites or other online services, you consent to the practices described in our Privacy Policy and Terms of Service, including our Cookie Policy.