Anatomy of a player card

Average career stats over an 84-game season: G, A, P, +/− and PIM are per-84 rates (so players with different career lengths compare directly); the TOI entries are averages per game (total, power play, short-handed). Under the bio line sits the player’s archetype — a k-means style cluster over the player’s last 100 GP (icon + name; 8 forward clusters plus the rule-based Playmaker, 7 defense types, era-adjusted features; the full icon glossary is on the Reference page). Skaters with 60–99 career GP are typed on all their games so far; below 60 there is no type. After it comes the steadiness tag (zigzag icon, band, percentile among active skaters at the position; higher = steadier): how often and how far the player’s games fall below average compared with players of the same net goals/84 and ice time. It’s a style trait, not value: shooters skew boom-or-bust because goals come in lumps. Shown from 100 career GP. The flag beside the name is the country he represents internationally, from EliteProspects (William Nylander is Sweden, though born in Calgary); hover it for a second nationality and his birthplace when they differ. A player EliteProspects hasn’t been matched for yet shows his country of birth, and the Draft line under the subtitle gives the draft year, the team that picked him, the round and the overall pick (or Undrafted). On the right of the header, the contract panel: his current cap hit with its share of this season’s cap, then one segment per season of the deal — dark for seasons done, the current season filled as far as the season has been played (the thin blue tick marks today), pale for seasons still to come, and hatched segments after a small gap for an extension he has already signed. Underneath is what happens when it runs out: UFA or RFA, the summer, and his age then. In the last year of a deal the current segment and a “final year” tag turn orange. Hover the panel for the signing date and total value, or a segment for its season. Contract terms come from PuckPedia and CapWages; a player whose new deal isn’t in either yet shows the cap hit alone.

One row per season, oldest at the top, with the team(s) he played for in the Team column (hover a logo for the abbreviation) and then the Box score group: his games played (GP), goals, assists, points, plus-minus, penalty minutes, power-play points (PPP), short-handed points (SHP), game-winning goals (GWG: the goal that put the winning team ahead for good; none in a shootout win), shots on goal (SOG), hits (HIT), blocked shots (BLK), takeaways (TK), giveaways (GV) and hits taken (HITT: times he was hit, from the play-by-play; blank before 2007-08, when it starts naming the player hit) that season (regular season; in the Career to date view, career totals through that season). The Type column is the player's archetype as of the end of that season (his trailing 100 GP; with 60–99 career GP, all his games so far; blank below 60; the season in progress shows last season's type until it is complete; seasons before 2010-11 are matched to the same types, which are fitted on 2010-11 onward): an icon every season, the full name and window on hover. In the Career to date view the column shows the type of his whole career through that season instead (every game up to then, not just the last 100), so the last season shows his career archetype. Skaters who have taken a shootout attempt get a Shootout column, still inside the box score group (goals / attempts): shootout goals / attempts that season (regular season, 2010-11 on; in the Career to date view, career totals through that season; blank in a season without an attempt). OZ/DZ starts closes the group: his offensive-zone faceoff starts per defensive-zone start, where a start is a shift he began on a faceoff in that zone (all situations, regular season; neutral-zone starts are left out; league average about 1.06, so above 1 is sheltered toward offense and below 1 leans defensive; hover for the counts; Career to date: career totals through that season). Every box under Net goals per 84 games is the player's net goals added per 84 games vs a league-average player, in calibrated model units — blue adds goals, red costs them, on fixed scales (−42 → +42 for Overall, a tighter −12.6 → +12.6 for the components, so seasons and players compare directly); gray text under the EV offense / EV defense pair, PP and PK is TOI per game in that situation — 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); an empty cell = no rated value that season. Faint vertical rules separate the column groups. The Overall column, the last of the group (next to Elo), is outlined:

Did he do it against the other team’s top line or its fourth? A career summary of how the player’s 5-on-5 minutes went against weak and strong opposition. Every 5v5 second rates the opposing five by their average Overall net goals/84 that season (each regressed toward average by sample size, so a hot call-up doesn't make a line look elite; unrated skaters count as average). Each season those ratings are split into league-wide quarters by ice time: Q4 is the toughest quarter of five-man units anyone faced, Q1 the easiest (the gray line gives the typical opposing rating). Left bars: net goals per 84 above an average player at his position (forward or defenseman) facing the same quarter, from the grid’s own scoring added up over just those minutes — EV offense and defense, finishing on his own shots, and an even share of his Playmaking and Suppression (fitted as fixed player effects, so spread evenly over his ice time) — scaled to his career 5v5 ice time, as if all of it came against that quarter. Blue adds goals, red costs them, on the component rows’ ±12.6 scale; the comparison is with the same quarter because every skater does worse against Q4 than Q1. PP, PK, penalties and faceoffs are not part of it. Right bars: the share of his career 5v5 time spent against each quarter, split by venue — the dark upper bar is his home games, the light lower bar his road games (each out of his own home or road 5v5 time). The dashed line at 25% is an even split, so a top-pair defenseman leans to Q4 and a sheltered depth player to Q1; the home coach has the last change, so the gap between the two bars shows how much matchups are chosen for or against him. Hover a bar for minutes, the component split, on-ice xGF% and goals for–against. Only the career is shown: one season’s quarter is about 300 minutes and swings by ±7 goals/84. Read the levels rather than the shape: doing unusually well against Q4 relative to Q1 doesn’t carry over from season to season (year-to-year r ≈ 0). The player who rises to meet the top line makes a good story; it just isn’t a repeatable one.

The career's shape in one line: a trailing 84-game moving average of the Overall row, in the same net-goals-per-84 units. The x-axis is career games and is shared with the Career net goals line beside it, so the line starts blank until the 84th game. The dot marks the career peak, the number at the right end is the current value, and faint vertical lines are season boundaries. Even elite careers dip below zero here: 84 games is still a small sample, and puck luck takes its time evening out. Goalie cards draw the same line for GSAx/84 over the last 84 appearances.

The running career total of the same per-game net goals the form line averages: how much value the career has banked versus a league-average player, in goals. A flat stretch means league-average play, a rising slope means value being added; the red dot marks the career high-water mark when the current total has slipped below it. Goalie cards accumulate GSAx instead. On the website, the row of buttons above these two lines switches both of them from Overall to any one of the ten components (EV offense, EV defense, Power play, Penalty kill, Finishing, Playmaking, Suppression, Penalties, Faceoffs, Shootout); the ten add up to Overall game by game, and each component line is scaled to its own range. A second row, Box score, does the same for the counting stats (Points, Goals, Assists, PP points, SH points, Shots, Hits, Blocked shots, Takeaways, Giveaways, Hits taken, PIM, +/−): the form line becomes his total over the trailing 84 games and the career line his running regular-season total. The Games buttons under it add his playoff games to every view, net goals and box score alike (With playoffs): each spring's games slot in after that regular season and are drawn in amber. Playoff net goals are priced exactly like a regular-season game and measured against that season's regular-season average player, so a playoff stretch reads on the same scale as the games around it (the 84-game stretch bands are left off in this view, since they mark regular-season windows). On every view a dashed grey rule marks a change of team, labelled with the new team above the chart (his first team at the start; a franchise relocation is not a change). Its last button, OZ/DZ, plots the OZ/DZ starts ratio instead: pooled over the trailing 84 games, and career to date from his 20th game, with a dashed line at 1.00 (as many offensive- as defensive-zone starts).

Two windows of the 84-game form line above, shaded on it (B / W, the same colors as the swatches here): its peak (Best) and its low (Worst). The colored chip is his net goals/84 over those 84 games, on the Overall row's scale, with the first and last game dates under it (a stretch can cross seasons). The rest is the box-score line over the same 84 games: goals, assists, points, power-play points (goals and assists with his team a skater up; an extra attacker for a pulled goalie does not count), plus-minus, average ice time, penalties taken and drawn, faceoff win % (shown with 30+ draws), and his team's record in those games (W-L-OTL, overtime and shootout losses in OTL; 2003-04 adds ties). Skaters with 84+ career games.

Every regular-season game of his career in one chart: each bar is the share of his games that finished with a given net goals total for that single game (in goals, not per 84), red below zero and blue above, with the dashed line at his career mean and the share of games above and below zero at the right. The values are leverage-weighted: finishing, penalties and faceoffs count by how much the moment could swing the game, so a garbage-time goal at 7–1 counts for little (the on-ice rows are priced that way already). These are the same per-game numbers that split each result in the Elo column. Most players sit close to zero in a typical game. Stars aren’t told apart by their median game; they’re told apart by a long right tail of big nights. Games beyond the axis (−2 to +3.8 goals, 0.1-goal bins) are counted in the end bars; hover a bar for its share and game count. On the website, All / Home / Away limits the chart to those games, and a season button compares one season with the career: the career turns grey and the season is drawn as blue outlines, each as a share of its own games, in wider 0.2-goal bins, with the season mean dotted. Player home/road splits are mostly noise from one season to the next, so read the Home/Away views as description. Skaters and goalies with 20+ games; goalie cards chart single-game GSAx (−5 to +8, 0.25 bins; 0.5 when comparing a season).

His official plus-minus (regular season, 2010-11 on), one row per season, split by the manpower on the ice when each goal was scored. Each colored chip is his share of the official number from that situation, blue for plus and red for minus, shaded on its own column’s scale, so a dark chip in a small column such as 5v3 can be a single goal. Even strength (5v5, 4v4, 3v3): every goal scored with him on the ice counts, for or against. Power play (5v4, 5v3, 4v3, his team a skater or two up): the official stat ignores power-play goals, so the only thing that can register is a shorthanded goal against, and these columns are never positive. Penalty kill (4v5, 3v5, 3v4): the mirror image, only a shorthanded goal for counts, so they are never negative. The 4v3 and 3v4 states are mostly overtime penalties. Goalie pulled: 6v5 is time on the ice with his own goalie on the bench for an extra attacker, usually chasing a game late (and a few delayed penalties); 5v6 is defending against the other team’s extra attacker, where his team’s empty-net goals land. Other is whatever is left of the official number, so each row adds up exactly: mostly empty-net goals against while his team had its goalie pulled on a power play (6-on-4), plus the odd goal the NHL shift charts miss. Goalie in is the official +/− minus the 6v5 and 5v6 columns: the plus-minus he earned with both goalies in net. Why bother? Because pulled-goalie time is under 2% of a skater’s minutes but about an eighth of the spread in plus-minus, and it carries no skill from year to year. It’s mostly a record of who the coach sends out: stars sent out to chase a game pile up 6v5 minuses (Ryan O’Reilly is −104 there for his career), while defensive defencemen trusted to close games collect 5v6 pluses. The Total row sums every season. Hover a cell for his minutes in that situation and the goals behind the number (on the power play and penalty kill it also shows the power-play goals the official stat leaves out). Skater cards only; the study behind it is on the Research page.

The table is his career shots on goal by type (2010-11 on). The bar is that type's share of his shots; the dark tick marks the league share for his position over the same seasons, each season weighted by his shots in it, so type-coding changes between eras cancel out. Then Total shots (shots on goal, goals included), Avg dist (average distance from the net in feet), xG (expected goals on those shots, the same shot model as Finishing), G, G/xG (goals per expected goal: 1.00 is average finishing; blank under 10 shots) and Net goals/84, the goals above xG those shots added per 84 games (gray under 100 shots). The Total row adds up every type. On the website the situation buttons switch the table as well as the map, and the shot-type buttons (All shots, Wrist, Snap, Slap, Backhand, Tip-in, Deflection, Other; or a click on a row) redraw the right half of the map for that type only, compared with the league's shots of the same type and situation (blocked shots carry no shot type, so the left half always shows all of them). Below: a full rink drawn from his side (he attacks left to right, his left on top). The right half is where his own shots on goal come from; the left half is where he blocks shots in his own zone. Each 4-ft zone is colored by its share of those shots minus the league average for his position in the same situation: blue = more of the shots come from there, red = fewer (±0.30 percentage points per zone). Small samples are pulled toward the league average: each zone’s difference is multiplied by n/(n+K), with K = 100 of his own shots and 180 blocks, set from how well odd-numbered games predict even-numbered ones — so a thin sample fades toward white instead of flashing noise. The block map is his own: with about 1,000 blocks, 86% of it repeats from one half of his games to the other. On the website, the All / Even strength / Power play / Penalty kill buttons redraw both halves for that situation from his point of view: on the penalty kill the right half is his shorthanded shots and the left half the power-play shots he blocks. A half with too few events in that situation (30 of his own shots, 25 blocks) stays blank. Goalie cards show a save map instead: each zone colored by save percentage above or below what that zone's shots predict.

Shown for players with at least 200 career draws (2010-11 on). The rink is drawn from his side: he attacks left to right, and his left is at the top, so each dot means the same thing on every card whichever end his team was defending. Each circle is his win percentage at that dot, colored by how far it sits above (blue) or below (red) the league’s win rate at the same dot (the small number, in percentage points; the scale runs to ±10). The bar under each dot is his share of all his draws taken there, with a tick at the league-average share. Both baselines are weighted by the seasons he played, so an early career and a recent one compare fairly. The line under the rink sums the shares by zone: a defensive specialist leans to his own zone, a top-line centre takes more of the centre-ice draws after goals and at the start of periods. Handedness is the biggest pattern in both halves. A left-shot centre wins about 54% on his left-side dots and about 44% on his right; right-shot centres lean the other way, less sharply, because most of the centres they face shoot left. Coaches use it: in recent seasons a right-shot centre takes about three-quarters of his end-zone draws on the right-side dots, so his bars will stand apart from the league ticks, which pool mostly left-shot centres. Since 2019-20 the attacking team chooses the dot after an icing, and left-shot centres’ share on the attacking-zone left dot rose from 19% to 26%. On the website, the All / Even strength / Power play / Penalty kill / 3v3 buttons redraw the rink for draws taken in that situation (by his team’s skater count, so a pulled-goalie extra attacker counts as the power play), with the league baselines taken from the same situation: power-play centres win more everywhere, and penalty-kill draws sit mostly in the own zone, so those two views stretch the share bars to 0–60%. 3v3 (regular-season overtime, almost all; also counted in Even strength) does the same: about 40% of 3v3 draws are at centre ice, one to open each overtime. A situation with fewer than 50 career draws shows a note instead. Players with at least 50 career games at centre also get a Head-to-head on the draw row under the rink: three small tables of his record against the five players he has taken the most faceoffs against (Most faced), and the five he has the highest and the lowest win% against among opponents with at least 25 draws. Every draw between the two since 2003-04 counts, regular season and playoffs, in all situations (the buttons do not filter it); win% is blue above 50% and red below; hover a row for the games and the playoff part.

Goalie cards read the same way with goalie panels: the header row shows career shot-quality totals — shots faced, GA, xGA, GSAx (goals saved above expected), Sv%, xSv% and Reb/100 vs lg (rebounds yielded per 100 shots faced, minus the league rate that season; 0 is average, lower is better — the raw rate roughly doubled between 2019 and 2025 through play-by-play recording changes, so it is only meaningful relative to its own season). The season table (one row per season, with games played, Elo and the season's cap hit, player value and contract value) shows GSAx per 84 games (including the shootout mirror beside it), a Shootout column — the win probability the shootout attempts he faced took away or saved, 3 goals per shootout win, as season totals (to date: the running total) — and a Reb/100 column in the same season-relative units; its Season | Career to date buttons work as on skater cards — Career to date shows his GSAx/84 over every appearance up to each season, regressed toward an average goalie, and the shots-faced-weighted Reb/100 through that season, so a single bad year reads against the career it sits in (display only: the season projections still regress toward an average goalie, which tested better one season ahead). Then the 84-appearance form line, the cumulative career GSAx line, the single-game GSAx histogram and the save map described above.