NHL franchises are slowly embracing the ‘Moneyball’ strategy of using analytics to assist in coaching, training and player evaluation but the league has been behind the other sports for two reasons. Hockey has more of a flow aspect to it and to build analytics models around the sport of hockey, you must have a very deep understanding of the game. This second has proven to be the biggest obstacle in hockey embracing analytics but a firm out of Park City, UT is changing this. The Sports Analytics Institute was founded in 2008 by two university professors that have extensive analytics backgrounds and a deep understanding and passion for hockey. Kevin Mongeon and Mike Boyle founded the company because they shared these unique skill sets and saw an opportunity for an unconventional sport to evolve through the implementation of analytics.
Mike and Kevin spoke with icrunchdata and discussed the uniqueness of hockey and how it compares, or more importantly doesn’t compare, to other sports and some of the challenges of implementing an analytics approach into a sport that is steeped on traditional methods of player evaluation and coaching. Kevin and Mike will also be speaking at the Predictive Analytics World Toronto event on May 12-15 and they explained what attendees should expect when they attend their case study.
You’ve been working with NHL teams since 2008 but it seems that hockey is behind in using analytics compared to baseball and football. What is the current state of NHL teams’ adoption of analytics into their organizations?
“Teams are moving forward with this new approach, some quicker than others, by hiring outside companies like ours and even hiring people within their organizations to explore this approach. The teams that are hiring internal analytics professionals are having a tough time finding people that are qualified in the area of analytics, hockey, and how to blend the two into actionable decisions for the manager. It’s difficult for them to write a job description and recruit for these people when it’s still unclear exactly what they will be doing. Using an outside organization like ours allows us to teach and train their internal staff on the areas that will provide the largest benefit from their analytics program.”
You were quoted in the article ‘Moneypuck’ and referenced your regression-type analysis that goes beyond straight stats and evaluates a player by his line mates, opponents, score, state of the game and other factors besides just their raw numbers. What are a few uncharacteristic variables that this analysis takes into consideration?
“Hockey is a very uncharacteristic sport in regards to the time of games, score of games, the incentives that change as the game changes and the fact that playoff games are potentially different from the regular season. In football, a player only plays offense or defense, basketball players play both offense and defense but it is a very high scoring game with typically a lot of lead changes but hockey is better compared to soccer. Low scoring, very complex and it comes down to a lot of small on-ice and off-ice decisions that make the difference. With the data we have for hockey, it requires a high degree of analytical skill to turn it into information that allows the manager to make better decisions.”
You are speaking at the Predictive Analytics World Toronto event on May 12-15th and your case study is titled ‘Designing Effective Hockey Teams through Physical Diversity’. How does physical diversity impact the success of hockey teams?
“Results suggest that managers that employ a less and more diverse group of players in terms of height and weight respectively have a greater likelihood of winning games. However, physical diversity does not affect game-goal outcomes. Therefore, managers should consider the effects of diversity when designing team rosters, but it is not necessary for coaches to alter within game line-ups to account for their effects.”
“The following example illustrates the results in terms of their impact on a team’s winning percentage. Teams that are among the top compared to the bottom of the league in diversity of team height will have a lower winning percentage of approximately 0.05 points, and teams that are among the top compared to the bottom of the league in diversity of team weight will have a higher winning percentage of approximately 0.07 points, after accounting for other team quality factors.”
What is the main takeaway that you would like the audience to have after attending your case study?
“There is a direct correlation between sports and business. Use sports as a context to learn from and teach analytics and test economic phenomenon but remember that the goal in using analytics is not just for individual player evaluation but to build a successful team, just like in business.”
Join Mike, Kevin and other top analytics professionals from Abbott Analytics, Blackberry, Elder Research, Boire Filler Group, Fifth Third Bank and more at the Predictive Analytics World Toronto event on May 12-15.
By: Todd Nevins, Director of News & Media, icrunchdata & Writer & Sr Editor, icrunchdata news
Originally published at news.icrunchdata.com
