July 19, 2021 in Sports Analytics
Optimizer for the 2021 NHL Expansion Draft
Team from University of Toronto builds website that runs an integer programming model to play “armchair general manager.”
SHARE: PRINT ARTICLE:
https://doi.org/10.1287/orms.2021.04.17n
The National Hockey League (NHL), a professional ice hockey league in North America, is widely considered the top professional league in the world. This year, the NHL will expand to create a 32nd team, the Seattle Kraken. On July 21, an “expansion draft” will be held in which Seattle will build its inaugural roster of players by drafting from all of the other teams, except the (Las) Vegas Golden Knights (itself an expansion team from 2017 and exempt from this draft).
Draft Rules
Each team participating in the draft must follow a specific set of rules. For example, teams are allowed to protect a certain number of players in each position, which prevents Seattle from selecting those players. Teams must also expose a certain number of players for Seattle to select. Seattle, on the other hand, is allowed to select one player from each participating team, and the total annual contract value of those players must satisfy salary cap constraints. Current teams must carefully choose which players they expose. In the previous expansion draft, Vegas was able to construct a strong roster and ended up going all the way to the Stanley Cup Finals in its inaugural season, something that had never been accomplished before.
The rules that govern the expansion draft can naturally be formulated in an integer programming model. In particular, the rules form hard constraints, but each team’s objective function is up to them to decide. One can think of the objective function for each team as maximizing the “value” associated with the roster they keep/build. This value can be measured with several different metrics. For example, each player can be assigned a score based on how much they contributed to their team’s wins over the last season (i.e., value measures on-ice performance) or their future salary (i.e., value measures future financial flexibility). The integer programming model for existing teams minimizes the value of their exposed players, while Seattle’s model maximizes the value of their assembled team.
Expansion Draft Optimizer
Our team of engineers and graduate students from the University of Toronto have built an Expansion Draft Optimizer that lets users play “armchair general manager” for Seattle or any of the current teams. Users first select a team and scoring function based on what metrics they think the team should optimize. The website runs an integer programming model with this objective function, showing the optimal projections that the selected team should make and optimal selections that Seattle should make. Afterward, users can vary the objective function for each team to see how the results change. They can also investigate the impact of manually selecting certain players to be protected or exposed.
After the success of Vegas in 2017, all eyes are on the upcoming draft. We believe the Expansion Draft Optimizer is an opportunity to showcase the power of operations research (O.R.) in a fun and accessible setting that would appeal to both sports fans and O.R. enthusiasts.
To use this model, visit our website at: https://www.nhlexpansiondraft.com. If you have any questions or comments, you can email us at [email protected] or find us on Twitter @NHLExpansionOpt.
Michael Shin is a software engineer at Intel. He studied at the University of Toronto and wrote his undergraduate thesis under the supervision of Prof. Timothy Chan. Yusuf Shalaby is a technology specialist at BAI Communications, leveraging cloud technologies to deliver data solutions to transit agencies. In 2020, he completed his MASc in operations research at the University of Toronto. Albert Loa is a senior O.R. product developer for the AD OPT team at IBS Software Canada, building products for airline crew planning optimization. In 2022, he completed his Master of Applied Science in operations research at the University of Toronto. Ben Potter is an applied research scientist in the Advance Technology Group at ServiceNow. In 2018, he completed a master’s in operations research at the University of Toronto. Timothy C. Y. Chan is the associate vice president and vice provost, strategic initiatives; Canada Research Chair in novel optimization and analytics in health; and professor in the Department of Mechanical & Industrial Engineering at the University of Toronto. His primary research interests are in operations research, optimization and applied machine learning, with applications in healthcare, medicine, sustainability and sports. He is a member of INFORMS. Rafid Mahmood is an AI resident in the NVIDIA Toronto Artificial Intelligence Lab. In 2020, he completed his Ph.D. in operations research at the University of Toronto. He is a member of INFORMS.
