WILD PAIR ANALYTICS WITH COACHES' INSTINCTS TO GET MOST OUT OF PLAYERSNov 21, 2018
By Tom Witosky
Like all sports, the outcome of a hockey game is determined simply – the team with the most goals wins.
But among National Hockey League teams and a growing number of American Hockey League teams, figuring out how to achieve that outcome has become the province not only of coaches but statistical analysts like Alexandra Mandrycky, an Atlanta, GA native and hockey operations analyst for the Minnesota Wild since 2016.
“We look for those things that we can use to predict scoring,” said Mandrycky, who also consults with Iowa Wild on data issues, while on a recent visit to Des Moines. “Obviously, the goal of the game is to score more goals, but what we try to do is figure out how each player on our teams can do that more often.”
In some hockey circles, the term “analytics” is derided by those who believe statistical data analysis robs the game of its tradition, verve and color – when crunching bodies in the corners was more important than crunching numbers with a computer.
Others, like Keith Paulsen, the Iowa Wild’s video coach, think differently. Instead, they believe it is a helpful tool that provides vital information to be used in improving individual player and team performance over a season.
“For an everyday fan, analytics can be pretty complicated,” Paulsen said. “Frankly, for a coach it can be complicated as well. But, the people who do it are brilliant and the way they come up with things to help a team based on position, where the puck is on the ice, and where shots are taken is beyond brilliant.”
Shep Harder, assistant general manager for the Minnesota Wild, described the use of analytics as a tool to broaden the perspective of coaches and front office staff not as a final word on a player’s value or a team’s performance.
“I am a firm believer that it extends your reach and gives you data points, but at the end of the day it is supplemental or complementary information to what your scouts have, what your coaches have,” Harder said. “But at least it spurs discussion, conversation and gets you thinking.”
For example, Harder said, Minnesota defenseman Matt Dumba, who recently signed a long-term contract with Minnesota, provided a statistical foundation from the beginning of his career about his potential value.
“Going back to the 2005-06 season, Matt was one of five defensemen under the age of 23 to score 10 goals or more in three consecutive seasons. It is a simple question to ask, but a hard one to answer if you don’t have the data,” he said.
Just what are “analytics”?
Paulsen recently shared a report of an Iowa match in which computer analysis broke down the game into an eight-page document with ten categories including expected goals, shooting, creating scoring chances, power play, penalty kill, goaltending, possession time, and puck management.
Within each category, the report provided statistical information on nearly 200 measurable aspects of the game. Those aspects range from measuring the success rate of goals scored when expected to shots on net off of a cycle, a rush, a forecheck or a faceoff when at even strength, on a power play, or on a penalty kill.
Each aspect is measured for every player and compiled during the season to provide an accurate picture of a player’s strengths and weaknesses.
“We can now know how much time the Kloos line was on the ice against a particular line,” Paulsen said of Iowa’s line made-up usually of Justin Kloos, Sam Anas and Kyle Rau. “They will be able to tell us the success rate that the line had related to shots for and shots against, on goal, goals scored. It can be broken down to tell you how the Kloos line did against all four of the opponent’s lines as well as all three of the d-pairings.”
In addition, the data is compiled and analyzed for each opponent. It’s also provided within just a day or two of the game.
“What is so valuable is that when you play a team like Rockford or Milwaukee ten times a year or a Chicago eight times in a year, those are important things to know,” Paulsen said. “You think you know it as a coach, but you don’t always know. That just solidifies what you are thinking.”
All of the analysis begins with tracking how much time each player is on the ice. To get accurate accounting of time on ice, Paulsen has a group of Drake University students monitor players coming on and off the ice. It’s usually a team of three or four interns sitting in the Wells Fargo Arena press area compiling the information.
“It is as simple as tracking time on ice,” he said. “When I get that information, time on ice and shots is what I provide them and then they come up with a lot of stuff.”
Mandrycky and Andrew Thomas, the Wild’s leading data researcher, are the ones responsible for coming up with “a lot of stuff.”
Chuck Fletcher, then Minnesota’s general manager, hired the two analysts who had attracted attention through Thomas’ development of the website War on Ice, which tracked statistical information from NHL games. Mandrycky, a Georgia Tech graduate, had joined the staff initially as a volunteer.
“I didn’t start watching hockey until I was 17,” she said. “I had a few friends who were into it and they got me to watch. I would watch football but then I discovered hockey was a really good way to spend three hours.”
Mandrycky said her major goal is to make sure the information is useful and understood.
“In general, what I try to do is make sure they understand what it is they are reviewing,” she said. “In any job, information can be overwhelming and information by itself may not be very valuable unless there is understanding of what is behind it.”
As a result, she makes sure the information “has a point and we just aren’t shoving data into their faces because that can be scary.”
Paulsen said the statistical information has allowed coaches to follow their instincts much of the time.
“It gives a coaching staff information that allows us to make changes and simply be more flexible if it’s determined that things aren’t working as they should,” he said.
Harder said that is the essence of making progress whether it is on the ice or anywhere else.
“No matter the industry, anytime you have something that can challenge your thought process, make you think about things, and lets you see the domino effect of things, that is a positive and an asset,” he said. “We are big believers in analytics, but it isn’t a mathematical construct in one number that defines a player’s value. I don’t believe in that. It has to be within a context of how a coach uses or deploys a player within the system.”