Game Score (GS)
A single-number summary of everything a skater did in one game -- goals, assists, shot quality, play-driving, penalties and the little things -- on a scale where about 0.1 is a typical night and 2.0 is a star performance.
Game Score answers the question a box score can't: who was actually the best player tonight? Points alone are noisy over sixty minutes -- a defenseman can dominate territorially and finish with nothing beside their name, while a forward can tap in a rebound and lead the scoresheet. Game Score rolls a skater's whole night into one number, so a shutdown defenseman and a goal-scoring winger can be compared on the same scale.
Hockey Alchemy uses Dom Luszczyszyn's Game Score with his 2019 revisions. The original 2016 version leaned on shot attempts (Corsi); the modern one replaces raw shots with individual expected goals and raw Corsi with on-ice expected-goal differential, so a slot chance counts for more than a point shot. Weights are position-specific: defensemen get more credit for driving play (0.8 on xG differential versus 0.5 for forwards), because that is more of their job.
Game Score is a single-game descriptive measure, not a projection. It says what happened, not what will happen -- one huge night doesn't make a player elite, and Game Score makes no attempt to adjust for opponent or usage. For a season-long, replacement-adjusted view of value, use GAR instead. The two answer different questions: Game Score rates a night, GAR rates a career.
Formula
Game Score = 0.75*G + 0.7*A1 + 0.55*A2 + 0.075*ixG
+ 0.05*BLK + 0.15*PenDrawn - 0.15*PenTaken
+ 0.01*(FO won - FO lost)
+ 0.05*Takeaways - 0.05*Giveaways + 0.05*Hits
+ on-ice differentials (for - against), by position:
Forwards: 0.5*xG+/- + 0.25*Goals+/-
Defensemen: 0.8*xG+/- + 0.30*Goals+/-
+ special teams:
0.12*xG+/-(PP) + 0.20*Goals+/-(PP)
0.12*xG+/-(PK) + 0.20*Goals+/-(PK)Components
| Component | Description |
|---|---|
| Individual production | Goals (0.75), primary assists (0.7) and secondary assists (0.55) -- primaries are worth more because they are more repeatable than secondaries. |
| Individual xG | The quality of a skater's own shots (0.075 each), replacing the raw shot count used in the 2016 formula. A slot chance counts for more than a point shot. |
| Play-driving | On-ice expected-goal and goal differential while the skater is on the ice, weighted more heavily for defensemen (0.8) than forwards (0.5). |
| Special teams | Power-play and penalty-kill differentials, weighted separately from even strength. |
| The little things | Blocks, hits, takeaways, giveaways, faceoff differential and penalties drawn or taken -- small weights that mostly break ties. |
Frequently Asked Questions
What is a good Game Score?
Across 47,195 skater games in 2025-26, the median was 0.10 and the average 0.34 -- a typical night barely moves the needle. About 9.4% of games cleared 1.0, only 1.3% cleared 2.0, and 0.1% cleared 3.0. So 1.0 is a good night, 2.0 is a star performance, and 3.0+ is one of the best games anyone plays all season.
What is the highest Game Score this season?
In 2025-26 the best single game belonged to Evan Bouchard at 4.27 (3 goals, 3 assists), followed by David Pastrnak at 3.94 on six assists and Logan Cooley at 3.79 on four goals. Note that Pastrnak scored none himself -- Game Score rewards the whole performance, not just the goals.
How is Game Score different from GAR?
Game Score rates one game and is purely descriptive; GAR rates a season and is measured against replacement level, adjusted for teammates and competition via RAPM. Use Game Score to find the best player in a given game, and GAR to find the best player over a year.
Can goalies have a Game Score?
This formula is for skaters only. Goaltenders are evaluated with Goals Saved Above Expected (GSAx), which compares the goals they allowed to the expected-goal value of the shots they faced.
Why do defensemen and forwards use different weights?
Because their jobs differ. A defenseman who tilts the ice without producing points is doing their job well, so on-ice xG differential carries 0.8 for defensemen against 0.5 for forwards. Without that split, defensemen would be systematically undervalued by a points-heavy metric.