Player Card — Annotated Guide

The numbered markers explain each element; every card in the net-goals decks (player_cards_ng82.html and the team decks) reads the same way.

Nathan MacKinnon

C · #29 · COL

Average career stats over an 82-game season 1

+20

+/−

PIM

20:40

TOI/GP

3:35

PP TOI/GP

0:16

SH TOI/GP

$12.60M

CAP HIT 2026-27

12.1%

SHARE OF TEAM CAP

+0.73

NET GOALS/82 PER $M 3

The numbered elements

Basic stats

Average career stats over an 82-game season: G, A, P, +/− and PIM are per-82 rates (so players with different career lengths compare directly); the TOI entries are averages per game (total, power play, short-handed).

Net goals grid

One column per season ('11 = 2010-11) with the team(s) played for under each header; the CAREER column is the games-played-weighted mean. Every box is the player's net goals added per 82 games vs a league-average player, in calibrated model units — blue adds goals, red costs them, on fixed scales (−41 → +41 for the Overall row, a tighter −12.3 → +12.3 for the component rows, so seasons and players compare directly); gray text under a row is TOI per game in that situation (GP for the Overall row) — situation here means skater advantage: 6v5 with the goalie pulled counts as PP time and defending it counts as PK time, so these splits run a little higher than the official PP/SH TOI in the header row (the totals agree); a plain gray box = fewer than 10 GP that season. The Overall row and the CAREER column are set in bold with outlined boxes:

Overall — the sum of the seven component rows below; the CAREER box is the numerator of the contract-efficiency tile.

EV offense — on-ice expected-goal creation at even strength vs league-average context (score, time, zone starts), scaled to its calibrated marginal value — five skaters share the same flow, so the marginal is ~4.4× smaller than face value.

EV defense — the same flow suppressed: on-ice expected goals against vs context, same marginal scaling.

Power play / Penalty kill — priced chance flow above (PP) or suppressed below (PK) league-average context in situation, same marginal scaling.

Finishing — goals above expected (xG) on the shooter's own shots on goal — individual credit, at face value.

Penalties — penalties drawn − taken, at 0.146 goals per net penalty (the league's mechanical PP-conversion value), weighted as in the season simulator.

Faceoffs — net faceoff wins × 0.010 goals (fixed-effects calibration). Shown for context: faceoff value flows through possession, so it partly overlaps the on-ice rows and is not part of the headline rating.

Team value

Contract efficiency: career net goals added per 82 games, per $M of cap hit — the grid's CAREER Overall box divided by the cap hit, so the tile and the grid always agree (goalies: GSAx/82 per $M).

Glossary of statistical terms

xG (expected goals) — the probability that a given shot becomes a goal, based on its location, type, and game situation. Game situation is the manpower state when the shot is taken (even strength, power play, short-handed, or empty net). The same shot from the same spot converts far more often with a man advantage than it does 5 on 5; rebound shots (within 3 seconds of a prior shot) and rush shots are also priced separately. A rush is inferred from the play-by-play sequence: the attempt comes within 4 seconds of a faceoff, hit, giveaway, or takeaway that happened in the shooting team’s neutral or defensive zone — the puck was at the far end of the ice moments earlier, so the defense never got set. Rushes that start by winning a turnover (a takeaway, or an opponent’s giveaway) are priced apart from those following a hit or faceoff, since a true counter-rush converts very differently from a fling after a dump-in or neutral-zone draw. Summed over shots, xG says how many goals a shooter “should” have scored — or a goalie/team should have allowed (xGA).

WPA (win probability added) — a weight on each chance for how much it moves the odds of winning the game: a chance in a tied third period counts more than one in a blowout. “Priced” chance flow means xG weighted this way.

rel (relative to context) — measured against what a league-average player would produce in the same circumstances: score state, home/road, zone starts, and situation.

Elo — a rating updated game by game: it rises when the player outperforms expectation and falls when he doesn’t, so it rewards sustained performance over a hot stretch.

rel-xG-WPA Elo — the engine behind the on-ice rows (EV offense/defense, PP, PK): win-probability-priced expected-goal flow, measured relative to league-average context. A per-season Elo run on it drives the model’s validation and the 0-100 ratings edition of the cards.

Per 60 — a rate per 60 minutes of ice time in that situation, so heavy- and light-usage players compare fairly.

Deployment-adjusted — corrected for how the player is used (zone starts, score, time on ice) so sheltered minutes don’t inflate a score.

Ridge attribution / teammate-adjusted — a ridge regression over on-ice lineups that separates the player’s own contribution from that of the linemates he shares the ice with.

Exposure-weighted — an average weighted by playing time (games or minutes), so a 20-game season counts less than a full one.

GSAx (goals saved above expected) — xGA − GA: positive means the goalie stopped more than the quality of shots he faced predicted. xSv% is the save percentage those shots implied.

Reb/100 — rebounds yielded per 100 shots faced; lower is better.

Anatomy of a player card