The Line Blender: Olympic Lineups with Announced Rosters
2026-02-01
Using the announced Olympic rosters, we can apply our GNN model to the players who will actually compete in Milan-Cortina. Unlike the previous Olympic roster post, where we considered all eligible NHL players, here we're constrained by the national team rosters.
Using MoneyPuck data from both 2024 and 2025 seasons, I've generated optimal line combinations using both linear programming (LP) and greedy optimization methods for Canada, Finland, Sweden, and the United States1:
One point of interest that I didn't touch on in the series is that within position groups (forwards and defensemen), player positions were not further specified to the model. In other words, LD and RD are not differentiated, nor are LW, C, or RW. Nonetheless, the produced lines and pairings largely cover expected positions, suggesting that chemistry and playing style captured in the data naturally align with traditional positional roles.
It'll be interesting to see as the games are played how the actual lineups are constructed and how they compare to these model-generated combinations. I will be curious to see what players are played out of their usual position, with what linemates.
All posts in this series:
1. The Line Blender: Optimizing Lineups Using MoneyPuck's Expected Goals Percentage (xG%)
2. The Line Blender: Embedding Line Performance Using a GNN
3. The Line Blender: Using a GNN To Produce Olympic Rosters
4. The Line Blender: Using GNN Embeddings for Player Rankings
5. The Line Blender: Olympic Lineups with Announced Rosters
6. The Line Blender: Hypothetical Russian Olympic Lineups