Metrics

Goals scored. Own goals are excluded.

The final pass or touch before a goal, as the event record marks it.

Minutes on the pitch, from kickoff or entry until exit or full time.

Total shot attempts, including blocked and off target.

How many seasons of the selected window this player appears in.

Aerial duels won

Starts where the observed role matched the assigned depth

Corners taken

Total announced attendance across all matches that season (some matches are missing attendance data)

Blocks

Passes-into-the-box component

Clean sheets

Ball-carrying component

Clearances

Red cards

Yellow cards

Draws

Defensive VAEP

Defensive-value component

Corner deliveries

Regular-season finish

Fouls committed

From

Goals against

Goal difference

Goals for

Games played (regular season)

Games started

Total announced attendance at regular-season home matches (some matches are missing attendance data)

Interceptions

Key passes

Losses

Matches with signal coverage

Minute she came off — blank means she finished the match

Minute she entered — blank means she started

Penalty goals

Penalties faced

Penalty attempts

Penalties saved

Standard deviation of points-per-game across the table — LOWER means a tighter, more competitive league

The Shield winner's points (season lengths differ — see Teams)

Ball recoveries

Saves

Shots on target

Regular-season starts with an observed-role inference (outfield starters only)

Clubs in the league that season

Tackles

Tackles won

To

Seasons in league

Wins

Goals × 90 ÷ minutes played.

Shots on target ÷ total shots.

Aerial duel success rate

Assists per 90 minutes

Goals plus assists per 90 minutes

Goals against per 90 minutes

Share of deliveries producing a goal

Pass accuracy (excl. crosses & corners)

Pass events per 90 minutes, regular season (both teams)

Number of rated players that season

Points per game

Rank among all rated players that season

Saves ÷ (saves + goals against)

Share of deliveries producing a shot

Tackle success rate

Tackle events per 90 minutes, regular season (both teams)

Wins ÷ games played (draws not counted as half-wins)

The probability an average finisher scores each chance, summed. A model output.

Win, draw, and loss probabilities implied by compatible expected-goal totals, weighted as league points.

Expected goals with penalties removed.

Every action valued by how much it changed the chance of scoring and conceding.

How much a move raised the chance of scoring, by where it moved the ball.

The threat a defensive action destroyed.

The chance a shot scores given where it actually ended up on goal. A keeper metric.

Our overall player rating — five components combined into one number.

The role a player's actions most resembled — not the role they were listed at.

The finishing signal: goals scored above or below what the chances were worth.

Team xG for minus xG against — the run of play in one number.

Shot-stopping over expectation: saves worth more than an average keeper's.

Advanced 8 — assigned central mid, played like an attacking mid

Starts written at attacking mid (10)

Attacking wingback — assigned fullback, played like a winger

Game-state adaptability (higher better)

Big chances created per 90 (higher better)

Penalty-area entries per 100 possessions (higher better)

Starts written at center back (4/5)

Starts written at central mid (8)

Share of progression via carries (higher better)

Overall composite score — season average across all selection-model runs

Average calibrated confidence of the observed-role inference (0-1)

Lineup continuity (higher better)

Counter-attack threat conceded (lower better)

DAxT from interceptions

DAxT from tackles

Deep-lying midfielder — assigned central mid, played like a defensive mid

Starts written at defensive mid (6)

Dropping playmaker — assigned attacking mid, played like a defensive mid

Deep completions conceded rate (lower better)

Directness index (style, no better/worse)

Discipline risk (lower better)

Share of duels won (higher better)

False 9 — assigned striker, played like an attacking mid

Starts inferred as a fullback from her on-ball events (~77% accurate)

Fatigue load (lower better)

Recent-form adjustment in this week's run

Finishing over/under-performance

The model never classifies goalkeepers — column kept so the two blocks match

Half-back — assigned center back, played like a defensive mid

Inverted winger — assigned wide, played like a striker

Possessions disrupted per opponent possession (higher better)

Lead protection index (higher better)

Median seconds from loss to shot conceded (higher better)

Opponent-specific adjustment for this fixture

The model's XI pick order for this fixture (1 = first name on its team sheet)

Observed role, inferred from her on-ball events (~77% depth accuracy) — read as “played like,” never “is.” Blank = not inferred (GK, sub, or no footprint).

Press regain rate (higher better)

Progressive pass rate (higher better)

Median seconds from regain to own shot (lower better)

Set-piece threat conceded (lower better)

Set-piece threat created (higher better)

Starts written at striker (9)

Cosine similarity of style embeddings (1 = identical style)

Shots on target against (from PSxG model)

Substitution disruption (lower better)

Tempo index (style, no better/worse)

Starts where the observed role differed in depth from the assigned role

Share of starts with a role tension

Shot-triggered pG created per 90 (higher better)

Shot-triggered pG conceded per 90 (lower better)

Share of attacks from transitions (higher better)

Defensive VAEP

VAEP per 90

Wide playmaker — assigned wide, played like an attacking mid

Average pG per shot (higher better)

Model-predicted goal probability

Model-predicted shot probability per delivery

Expected goals per 90

Expected goals per shot (model shot count)

Expected goals against

Expected threat per 90