In game one of the Colorado–Minnesota series on April 17, the Avalanche were down 3-2 with three minutes left on the clock — so Colorado head coach Patrick Roy pulled his goalie. They scored once with the extra attacker before winning 4-3 in overtime.
If you follow hockey, you might not necessarily call the move shocking, but goalie-pulling is a tactic usually left to the last minute, or maybe the last minute and a half. However, SFU’s Tim Swartz, a professor in the Department of Statistics and Actuarial Science, who specializes in sports analytics, believes Roy’s strategy should be the norm.
In a paper co-authored with David Beaudoin, an assistant professor at Laval University, Swartz stated that a team down by one goal would do best by pulling their goalie with three minutes left.
While three minutes is a comfortable average, Swartz suggests that coaches should not restrict themselves to pulling their goalies at the end of the game. Instead, they should do so when the opportunity is ripe, such as when their team is awarded a power play and, thus, a man advantage. For example, as a junior coach, Roy once pulled a goalie with 17 minutes remaining in the game.
A team down by one goal would do best by pulling their goalie with three minutes left.
“It may or may not work out for you,” stated Swartz, explaining that pulling the goaltender does not guarantee success. “But what we’re interested in is the long-run properties [. . .] If you follow our strategies, we estimate you would be several points higher in the standings. That can be the difference between making the playoffs and not making the playoffs.”
Swartz also does research into the management side of the game, on topics such as drafting and player contracts, in addition to analysis of game strategies. He, along with graduate student Gerald Smith, is currently interested in developing a system of player evaluation. His system would be an improvement on the age old plus/minus measurement — a bane to many advocates of the use of advanced statistics in hockey.
One of the advantages of sports analytics is that it allows students to apply classroom lessons to real world problems. “You’re learning all of the statistical theory and methods but you’re applying them to sports problems,” said Swartz. He also pointed out that sports analytics have funnelled some of his students into jobs at hospitals and government organizations.
Aside from the accompanying career opportunities, Swartz admitted to The Peak that what is truly fulfilling about sports analytics comes from the love of the game. “I get lots of requests to supervise students who want to work on sports problems,” said Swartz. “This is a lot of fun working in this area.”
