MBB — additional Python functions — Models and calculators: AssistEvent–build_priors

AssistEvent

AssistEvent(player_code: 'str', count: 'ShotClockStats' = <factory>) -> None

One assist relationship's counts (LineupEventStats.AssistEvent,

LineupEventStats.scala:64-67).

Parameters

Parameter	Type	Default	Description
player_code	str		The other player in the assist event (by code).
count	ShotClockStats	<factory>	The assist counts, by shot-clock segment.

AssistInfo

AssistInfo(counts: 'ShotClockStats' = <factory>, target: 'Optional[list[AssistEvent]]' = None, source: 'Optional[list[AssistEvent]]' = None) -> None

Detailed assist info, split into given/received

(LineupEventStats.AssistInfo, LineupEventStats.scala:87-91).

Parameters

Parameter	Type	Default	Description
counts	ShotClockStats	<factory>	Raw assist statistics.
target	Optional[list[AssistEvent]]	None	Players "I" assisted, if tracked.
source	Optional[list[AssistEvent]]	None	Players who assisted "me", if tracked.

ConferenceId

ConferenceId(name: 'str') -> None

CBB conference identifier (ConferenceId, ``models/ConferenceId

.scala:7, AnyVal). **Scope addition, Task 5e.4** -- the first model consumed by mbb_ncaa_team_parsers.py (TeamIdParser.get_team_triples/build_lineup_cli_array/build_available_team_list``). Appended here (not inserted among the 5a-reviewed classes above) to keep this an additive-only change, matching RosterEntry's precedent.

ConferenceId.is_high_major (the companion object's other member, models/ConferenceId.scala:11-16) is NOT ported. It has no call site anywhere in TeamIdParser/TeamScheduleParser (verified: the only other ConferenceId construction sites in the upstream tree are kenpom/TeamParser.scala and BuildIngestPipeline.scala, neither of which is in this port's scope, and neither calls is_high_major either) -- nothing in Phase 5e would exercise it. Noted here rather than silently dropped, matching this module's precedent for other unreferenced companion-object members (see the module docstring's Year.until / Game.Score.by_winner notes).

Parameters

Parameter	Type	Default	Description
name	str		The unique name of the conference.

CutdownShotEvent

CutdownShotEvent(loc: 'Optional[ShotLocation]', geo: 'Optional[ShotGeo]', dist: 'Optional[float]', pts: 'int', value: 'int', is_ast: 'Optional[bool]', is_trans: 'Optional[bool]', is_orb: 'Optional[bool]') -> None

A narrowed ShotEvent, keeping only the fields needed once a

shot has been matched to a player/lineup event (CutdownShotEvent, models/ncaa/ShotEvent.scala:31-40). Scope addition, Task 5e.5 -- ported for shape fidelity even though it is dead code in the ENTIRE upstream tree: grepping shows it appears only in its own definition (ShotEvent.scala) and as the never-populated shot_info: Option[CutdownShotEvent] field on PlayerEvent.scala -- it is never constructed anywhere. (PlayByPlayUtils.shot_value is an UNRELATED event_str -> int point-value classifier that merely shares a similar name -- Task 5e.6 will NOT produce this type either.) Appended here (not inserted among the 5a-reviewed classes above) to keep this an additive-only change, matching RosterEntry/ConferenceId's precedent.

Parameters

Parameter	Type	Default	Description
loc	Optional[ShotLocation]		The shot's court location, in feet, if known.
geo	Optional[ShotGeo]		The shot's synthetic lat/lon, if known.
dist	Optional[float]		The shot's distance from the basket, in feet, if known.
pts	int		The point value if made (2/3), else 0.
value	int		The shot's attempt value (2/3), regardless of make/miss.
is_ast	Optional[bool]		Whether the shot was assisted, if known.
is_trans	Optional[bool]		Whether the shot was in transition, if known.
is_orb	Optional[bool]		Whether the shot followed an offensive rebound, if known.

Direction

Direction(*values)

Which team is in possession (RawGameEvent.Direction, :119-121).

FieldGoalStats

FieldGoalStats(attempts: 'ShotClockStats' = <factory>, made: 'ShotClockStats' = <factory>, ast: 'Optional[ShotClockStats]' = None) -> None

Field-goal counting stats (LineupEventStats.FieldGoalStats,

LineupEventStats.scala:75-79).

Parameters

Parameter	Type	Default	Description
attempts	ShotClockStats	<factory>	Shot attempts, successful or not.
made	ShotClockStats	<factory>	Successful shot attempts.
ast	Optional[ShotClockStats]	None	Successful shot attempts that were assisted, if tracked.

LeagueConstants

LeagueConstants(hfa: 'float', margin_sd: 'float', em_scale: 'float', avg_tempo: 'float', avg_efficiency: 'float', quad_thresholds: 'dict[str, dict[str, int]]', bubble_adj_em: 'float', in_game_wp_artifact: 'str') -> None

Per-league fitted constants for the prediction & tournament stack.

Algorithms in the stack are league-agnostic; every men's/women's-specific number lives here so a WBB caller is a by-reference shim plus this table (the same pattern wbb_rapm / wbb_ratings already use).

Parameters

Parameter	Type	Default	Description
hfa	float		Home-court advantage in points (fitted on the 2024 backtest).
margin_sd	float		Std. dev. of the game-margin residual (fitted on the 2024 backtest; the Brier-minimizing sigma agrees to within 0.04).
em_scale	float		Slope applied to the AdjEM difference when predicting a game margin. AdjEM is per-100-possessions, so a game margin scales by ~tempo/100 (~0.67); the fitted value is lower still because the as-of AdjEM estimate is noisy and the optimal predictive slope is attenuated (regression dilution). Fitted jointly with hfa.
avg_tempo	float		League baseline possessions per game (adjusted-tempo anchor).
avg_efficiency	float		League baseline points per 100 possessions.
quad_thresholds	dict[str, dict[str, int]]		NET-style quadrant opponent-rank upper bounds, keyed by venue (home / neutral / away) then q1 / q2 / q3 (Quad 4 is any opponent ranked worse than q3).
bubble_adj_em	float		AdjEM of a bubble-quality team on THIS engine's scale (mean of engine ranks 40-50 on the fit season) -- the WAB baseline.
in_game_wp_artifact	str		Filename of the bundled in-game-WP coefficients under sportsdataverse/mbb/models (fitted + committed in Phase 3).

LineupEvent

LineupEvent(date: 'datetime', location_type: 'LocationType', start_min: 'float', end_min: 'float', duration_mins: 'float', score_info: 'ScoreInfo', team: 'TeamSeasonId', opponent: 'TeamSeasonId', lineup_id: 'LineupId', players: 'list[PlayerCodeId]', players_in: 'list[PlayerCodeId]', players_out: 'list[PlayerCodeId]', raw_game_events: 'list[RawGameEvent]', team_stats: 'LineupEventStats', opponent_stats: 'LineupEventStats', player_count_error: 'Optional[int]' = None) -> None

A portion of a game during which a given lineup was on the floor

(LineupEvent, LineupEvent.scala:41-58).

Parameters

Parameter	Type	Default	Description
date	datetime		The date of the game.
location_type	LocationType		Home/away/neutral (etc.) for this game.
start_min	float		The point in the game at which the lineup entered.
end_min	float		The point in the game at which the lineup changed.
duration_mins	float		The duration of the lineup.
score_info	ScoreInfo		The score differential context for this event.
team	TeamSeasonId		The team under analysis.
opponent	TeamSeasonId		The opposing team.
lineup_id	LineupId		A string that defines the set of players on the floor.
players	list[PlayerCodeId]		Mapping from player code to full identity, for this lineup.
players_in	list[PlayerCodeId]		Players who subbed in for this event.
players_out	list[PlayerCodeId]		Players who subbed out for this event.
raw_game_events	list[RawGameEvent]		The raw NCAA event strings for both teams.
team_stats	LineupEventStats		Numerical stats extracted for the lineup (team side).
opponent_stats	LineupEventStats		Numerical stats extracted for the lineup (opponent side).
player_count_error	Optional[int]	None	If the lineup is "impossible", the number of players actually seen (for analysis purposes).

LineupEventStats

LineupEventStats(num_events: 'int' = 0, num_possessions: 'int' = 0, fg: 'FieldGoalStats' = <factory>, fg_rim: 'FieldGoalStats' = <factory>, fg_mid: 'FieldGoalStats' = <factory>, fg_2p: 'FieldGoalStats' = <factory>, fg_3p: 'FieldGoalStats' = <factory>, ft: 'FieldGoalStats' = <factory>, orb: 'Optional[ShotClockStats]' = None, drb: 'Optional[ShotClockStats]' = None, to: 'ShotClockStats' = <factory>, stl: 'Optional[ShotClockStats]' = None, blk: 'Optional[ShotClockStats]' = None, assist: 'Optional[ShotClockStats]' = None, ast_rim: 'Optional[AssistInfo]' = None, ast_mid: 'Optional[AssistInfo]' = None, ast_3p: 'Optional[AssistInfo]' = None, foul: 'Optional[ShotClockStats]' = None, player_shot_info: 'Optional[PlayerShotInfo]' = None, pts: 'int' = 0, plus_minus: 'int' = 0) -> None

A lineup event's full counting-stat tree (LineupEventStats,

LineupEventStats.scala:7-38).

Only num_events/num_possessions/pts/plus_minus are exercised by Phase 5a -- see the module docstring's scope note.

Parameters

Parameter	Type	Default	Description
num_events	int	0	Number of raw events folded into this lineup event.
num_possessions	int	0	Number of possessions attributed to this lineup event.
fg	FieldGoalStats	<factory>	Overall field-goal stats.
fg_rim	FieldGoalStats	<factory>	Rim field-goal stats.
fg_mid	FieldGoalStats	<factory>	Mid-range field-goal stats.
fg_2p	FieldGoalStats	<factory>	2pt field-goal stats.
fg_3p	FieldGoalStats	<factory>	3pt field-goal stats.
ft	FieldGoalStats	<factory>	Free-throw stats.
orb	Optional[ShotClockStats]	None	Offensive-rebound stats, if tracked.
drb	Optional[ShotClockStats]	None	Defensive-rebound stats, if tracked.
to	ShotClockStats	<factory>	Turnover stats.
stl	Optional[ShotClockStats]	None	Steal stats, if tracked.
blk	Optional[ShotClockStats]	None	Block stats, if tracked.
assist	Optional[ShotClockStats]	None	Assist stats, if tracked.
ast_rim	Optional[AssistInfo]	None	Rim-shot assist info, if tracked.
ast_mid	Optional[AssistInfo]	None	Mid-range-shot assist info, if tracked.
ast_3p	Optional[AssistInfo]	None	3pt-shot assist info, if tracked.
foul	Optional[ShotClockStats]	None	Foul stats, if tracked.
player_shot_info	Optional[PlayerShotInfo]	None	Per-player shot-quality info, if tracked.
pts	int	0	Points scored.
plus_minus	int	0	Point differential while this lineup was on the floor.

Methods

LineupEventStats.empty

LineupEventStats.empty() -> "'LineupEventStats'"

A fresh all-defaults LineupEventStats (:41).

LineupId

LineupId(value: 'str') -> None

The set of players on the floor, as an opaque id string

(LineupEvent.LineupId, LineupEvent.scala:172, AnyVal).

Parameters

Parameter	Type	Default	Description
value	str		The opaque lineup identifier.

LocationType

LocationType(*values)

Game location (Game.LocationType, Game.scala:36-38).

PlayerCodeId

PlayerCodeId(code: 'str', id: 'PlayerId', ncaa_id: 'Optional[str]' = None) -> None

A player's within-team-season code paired with their full identity

(LineupEvent.PlayerCodeId, LineupEvent.scala:185-189).

Parameters

Parameter	Type	Default	Description
code	str		The player code, unique within the team/season only.
id	PlayerId		The player's globally-unique identity.
ncaa_id	Optional[str]	None	The player's NCAA-issued id, if known.

PlayerEvent

PlayerEvent(player: 'PlayerCodeId', player_stats: 'LineupEventStats', date: 'datetime', location_type: 'LocationType', start_min: 'float', end_min: 'float', duration_mins: 'float', score_info: 'ScoreInfo', team: 'TeamSeasonId', opponent: 'TeamSeasonId', lineup_id: 'LineupId', players: 'list[PlayerCodeId]', players_in: 'list[PlayerCodeId]', players_out: 'list[PlayerCodeId]', raw_game_events: 'list[RawGameEvent]', team_stats: 'LineupEventStats', opponent_stats: 'LineupEventStats', player_count_error: 'Optional[int]' = None) -> None

A lineup event's stats, narrowed to one player (PlayerEvent,

models/ncaa/PlayerEvent.scala:48-70). Scope addition, Task 5c.4 -- deferred by 5a since only ~sportsdataverse.mbb .mbb_ncaa_lineup_enrich.create_player_events (5c.4) returns it. Appended here (not inserted among the 5a-reviewed classes above) to keep this an additive-only change.

Same field shape as LineupEvent with two fields prepended (player, player_stats) -- the Scala builds this via a shapeless.LabelledGeneric HList splice of PlayerEvent's own player/player_stats onto every field of a LineupEvent instance; this port has no generic-programming machinery, so ~sportsdataverse.mbb.mbb_ncaa_lineup_enrich.create_player_events constructs the dataclass directly instead.

SingleEventMeta / event_meta / game_id are NOT ported. PlayerEvent.scala's companion object nests a SingleEventMeta case class, but the two fields that would carry it (event_meta, game_id) are commented out in the Scala source itself (PlayerEvent.scala:67-69) -- never part of the live case class, and create_player_events never constructs a SingleEventMeta. Nothing to defer; there is no live field to port.

Parameters

Parameter	Type	Default	Description
player	PlayerCodeId		The player this narrowed event describes.
player_stats	LineupEventStats		The player's own numerical stats for this lineup event.
date	datetime		The date of the game.
location_type	LocationType		Home/away/neutral (etc.) for this game.
start_min	float		The point in the game at which the lineup entered.
end_min	float		The point in the game at which the lineup changed.
duration_mins	float		The duration of the lineup.
score_info	ScoreInfo		The score differential context for this event.
team	TeamSeasonId		The team under analysis.
opponent	TeamSeasonId		The opposing team.
lineup_id	LineupId		A string that defines the set of players on the floor.
players	list[PlayerCodeId]		Mapping from player code to full identity, for this lineup.
players_in	list[PlayerCodeId]		Players who subbed in for this event.
players_out	list[PlayerCodeId]		Players who subbed out for this event.
raw_game_events	list[RawGameEvent]		The raw NCAA event strings for both teams.
team_stats	LineupEventStats		Numerical stats extracted for the lineup (team side).
opponent_stats	LineupEventStats		Numerical stats extracted for the lineup (opponent side).
player_count_error	Optional[int]	None	If the lineup is "impossible", the number of players actually seen (for analysis purposes).

PlayerShotInfo

PlayerShotInfo(unknown_3pm: 'Optional[tuple[int, int, int, int, int]]' = None, early_3pa: 'Optional[tuple[int, int, int, int, int]]' = None, unast_3pm: 'Optional[tuple[int, int, int, int, int]]' = None, ast_3pm: 'Optional[tuple[int, int, int, int, int]]' = None) -> None

Per-player shot-quality info, keyed by lineup slot

(LineupEventStats.PlayerShotInfo, LineupEventStats.scala:98-103). Each tuple is a fixed-arity 5-slot (one per lineup spot), mirroring the Scala PlayerTuple[Int] = Tuple5[Int, Int, Int, Int, Int] alias.

Parameters

Parameter	Type	Default	Description
unknown_3pm	Optional[tuple[int, int, int, int, int]]	None	3pt makes of unknown assist status, per slot.
early_3pa	Optional[tuple[int, int, int, int, int]]	None	Early-shot-clock 3pt attempts, per slot.
unast_3pm	Optional[tuple[int, int, int, int, int]]	None	Unassisted 3pt makes, per slot.
ast_3pm	Optional[tuple[int, int, int, int, int]]	None	Assisted 3pt makes, per slot.

PlayerValueConstants

PlayerValueConstants(pace_baseline: 'float', bubble_recruit_rank: 'int', bundle_prefix: 'str') -> None

Per-league constants for the player-value spine.

Parameters

Parameter	Type	Default	Description
pace_baseline	float		League baseline possessions per game (per-100 scaling).
bubble_recruit_rank	int		National recruit rank of a "bubble" high-major rotation player (recruiting-model reference point).
bundle_prefix	str		Artifact filename prefix under mbb/models ("mbb" / "wbb").

PossCalcFragment

PossCalcFragment(shots_made_or_missed: 'int' = 0, liveball_orbs: 'int' = 0, actual_deadball_orbs: 'int' = 0, ft_events: 'int' = 0, ignored_and_ones: 'int' = 0, bad_fouls: 'int' = 0, offsetting_bad_fouls: 'int' = 0, turnovers: 'int' = 0) -> None

Running stats needed to calculate possessions for one lineup event,

one direction at a time (PossessionUtils.PossCalcFragment, PossessionUtils.scala:124-144).

Parameters

Parameter	Type	Default	Description
shots_made_or_missed	int	0	Count of shot attempts (made or missed).
liveball_orbs	int	0	Count of live-ball offensive rebounds.
actual_deadball_orbs	int	0	Count of dead-ball offensive rebounds.
ft_events	int	0	Count of free-throw sets (capped-at-1 flag per set).
ignored_and_ones	int	0	Count of and-one free throws ignored for possession purposes (capped-at-1 flag).
bad_fouls	int	0	Count of technical/flagrant fouls counted against the defending side (capped-at-1 flag).
offsetting_bad_fouls	int	0	Count of technical/flagrant fouls that offset (net zero) rather than counting against either side (capped-at-1 flag).
turnovers	int	0	Count of turnovers.

PossessionEvent

PossessionEvent(dir: 'Direction') -> None

Decomposes RawGameEvent\ s into attacking/defending sides

(RawGameEvent.PossessionEvent, LineupEvent.scala:126-149).

Parameters

Parameter	Type	Default	Description
dir	Direction		Which team (Direction.TEAM / Direction.OPPONENT) is currently in possession.

Methods

PossessionEvent.attacking_team

PossessionEvent.attacking_team(ev: 'RawGameEvent') -> 'Optional[str]'

The event string for the team in possession, or None.

Parameters

Parameter	Type	Default	Description
ev	RawGameEvent		The raw game event to inspect.

Returns

ev.team if dir is Direction.TEAM, ev.opponent if Direction.OPPONENT, else None.

PossessionEvent.defending_team

PossessionEvent.defending_team(ev: 'RawGameEvent') -> 'Optional[str]'

The event string for the team NOT in possession, or None.

Parameters

Parameter	Type	Default	Description
ev	RawGameEvent		The raw game event to inspect.

Returns

ev.team if dir is Direction.OPPONENT, ev.opponent if Direction.TEAM, else None.

RapmConfig

RapmConfig(...)

Port of RapmConfig (RapmUtils.ts:175-179).

RapmPlayerContext

RapmPlayerContext(...)

Port of RapmPlayerContext (RapmUtils.ts:147-173).

See the module docstring for why filtered_lineups is a Python callable rather than a materialized dict.

RapmPreProcDiagnostics

RapmPreProcDiagnostics(...)

Port of RapmPreProcDiagnostics (RapmUtils.ts:187-194) -- the

multi-collinearity diagnostic calc_collinearity_diag returns.

RapmPriorInfo

RapmPriorInfo(...)

Port of RapmPriorInfo (RapmUtils.ts:124-133).

RapmProcessingInputs

RapmProcessingInputs(...)

Port of RapmProcessingInputs (RapmUtils.ts:196-203).

See the module docstring's "Task 3.5 notes" for why soln_matrix and sd_rapm are plain nested lists rather than NDArrays, and why sd_rapm exists at all (a Python-only addition beyond upstream's own return shape).

RawGameEvent

RawGameEvent(min: 'float', team: 'Optional[str]' = None, opponent: 'Optional[str]' = None) -> None

A single NCAA play-by-play event line (LineupEvent.RawGameEvent,

LineupEvent.scala:65-105).

Exactly one of team / opponent is populated per event -- the raw string is the literal "date,time,event" line from the NCAA website.

Parameters

Parameter	Type	Default	Description
min	float		The game-clock minute (fractional) this event occurred at.
team	Optional[str]	None	The raw event string, if this event belongs to the team under analysis.
opponent	Optional[str]	None	The raw event string, if this event belongs to the opponent.

Methods

RawGameEvent.for_opponent

RawGameEvent.for_opponent(s: 'str', min: 'float') -> "'RawGameEvent'"

Build an opponent-side event (Scala ``RawGameEvent.opponent(s,

min), LineupEvent.scala:109-110-- renamed per the "Scala idiom decisions" module note to avoid colliding with theopponent`` field).

Parameters

Parameter	Type	Default	Description
s	str
min	float

RawGameEvent.for_team

RawGameEvent.for_team(s: 'str', min: 'float') -> "'RawGameEvent'"

Build a team-side event (Scala RawGameEvent.team(s, min),

LineupEvent.scala:107-108 -- renamed per the "Scala idiom decisions" module note to avoid colliding with the team field).

Parameters

Parameter	Type	Default	Description
s	str
min	float

RosterEntry

RosterEntry(player_code_id: 'PlayerCodeId', number: 'str', pos: 'str', height: 'str', height_in: 'Optional[int]', year_class: 'str', gp: 'int', origin: 'Optional[str]', role: 'Optional[str]') -> None

An entry in an NCAA team roster (RosterEntry, ``models/ncaa

/RosterEntry.scala:11-21). **Scope addition, Task 5e.1** -- the first model consumed by the HTML-parser layer (mbb_ncaa_roster_parser.py``). Appended here (not inserted among the 5a-reviewed classes above) to keep this an additive-only change, matching PlayerEvent's precedent.

The trailing role field (present in the Scala case class shape but never populated by RosterParser.parse_roster -- every construction site there, both the real-player and coach__branches, passes the literalNone for it) is presumably set by a later phase's box-score parser (BoxscoreParser`, out of this task's scope); it is carried here for shape fidelity even though Task 5e.1's only producer never populates it.

Parameters

Parameter	Type	Default	Description
player_code_id	PlayerCodeId		The player's code + full identity (the roster-row equivalent of a box-score/PbP player reference).
number	str		The jersey number, as printed (may be non-numeric text).
pos	str		The listed position.
height	str		The listed height, in "FT-IN" text form (e.g. "6-3").
height_in	Optional[int]		height parsed to total inches, if it matched height_regex.
year_class	str		The listed academic year ("Fr"/"So"/"Jr"/ "Sr"/etc.).
gp	int		Games played.
origin	Optional[str]		The player's hometown/prior-school text, if the source table has that column (v1 rosters only).
role	Optional[str]		Reserved for a later phase; always None from parse_roster (see above).

ScoreInfo

ScoreInfo(start: 'Score', end: 'Score', start_diff: 'int', end_diff: 'int') -> None

Score context at the start/end of a lineup event

(LineupEvent.ScoreInfo, LineupEvent.scala:153-158).

Parameters

Parameter	Type	Default	Description
start	Score		Score at the start of the event.
end	Score		Score at the end of the event.
start_diff	int		Score differential (team - opponent) at the start.
end_diff	int		Score differential (team - opponent) at the end.

Methods

ScoreInfo.empty

ScoreInfo.empty() -> "'ScoreInfo'"

A fresh zeroed ScoreInfo (ScoreInfo.empty, :161-166).

ShotClockStats

ShotClockStats(total: 'int' = 0, early: 'Optional[int]' = None, mid: 'Optional[int]' = None, late: 'Optional[int]' = None, orb: 'Optional[int]' = None) -> None

Counting stats broken down by shot-clock segment

(LineupEventStats.ShotClockStats, LineupEventStats.scala:51-57).

Parameters

Parameter	Type	Default	Description
total	int	0	Count across the entire shot clock.
early	Optional[int]	None	Count in the first 10s, if tracked.
mid	Optional[int]	None	Count in the middle 10s, if tracked.
late	Optional[int]	None	Count in the last 10s, if tracked.
orb	Optional[int]	None	Count in the first 10s following an offensive rebound, if tracked (else folded into mid/late as normal).

ShotEvent

ShotEvent(player: 'Optional[PlayerCodeId]', date: 'datetime', location_type: 'LocationType', team: 'TeamSeasonId', opponent: 'TeamSeasonId', is_off: 'bool', lineup_id: 'Optional[LineupId]', players: 'list[PlayerCodeId]', score: 'Score', min: 'float', loc: 'ShotLocation', geo: 'ShotGeo', dist: 'float', pts: 'int', value: 'int', ast_by: 'Optional[PlayerCodeId]', is_ast: 'Optional[bool]', is_trans: 'Optional[bool]', raw_event: 'Optional[str]') -> None

Info about one shot taken during a game, all distances in feet

(ShotEvent, models/ncaa/ShotEvent.scala:9-29). Scope addition, Task 5e.5 -- the model produced by ~sportsdataverse.mbb.mbb_ncaa_shot_parser.create_shot_event_data.

Fields mbb_ncaa_shot_parser.create_shot_event_data (this task) actually populates: player (best-effort -- tidy-resolved + coded for the team under analysis, name-coded only for the opponent), date/location_type/team/opponent (copied from the box-score lineup), is_off, score (re-oriented for home/away/ neutral perspective), min (ascending game-clock time, after phase1_shot_event_enrichment), loc/dist (transformed court coordinates + Euclidean distance from the basket, after the self-correcting flip pass), geo (synthetic lat/lon), raw_event (the SVG <title> text, for debugging).

Fields left as placeholders for a LATER phase (Task 5e.6, PlayByPlayUtils/ShotEnrichmentUtils): lineup_id (always None here -- "discard if bad lineup" per the Scala comment, filled in once the shot is matched against an actual on-floor lineup), players (always [] here -- filled in from the matched lineup), pts (not the real point value -- this task only sets it to 1/0 for made/missed, matching the Scala's own // (enrich in final phase) comment; the real 2pt/3pt value comes from PlayByPlayUtils.shot_value in Task 5e.6), value (always 0 here, same "final phase" note), ast_by/is_ast/is_trans (always None here -- assist/ transition attribution needs the play-by-play cross-reference Task 5e.6 builds).

Parameters

Parameter	Type	Default	Description
player	Optional[PlayerCodeId]		The shooting player's code + identity, if resolved (None is never actually produced by this task's parser, but the type allows for it per the Scala Option).
date	datetime		The date of the game.
location_type	LocationType		Home/away/neutral (etc.) for this game.
team	TeamSeasonId		The team under analysis.
opponent	TeamSeasonId		The opposing team.
is_off	bool		Whether the team under analysis is the one shooting.
lineup_id	Optional[LineupId]		The on-floor lineup id, if/when matched (see above).
players	list[PlayerCodeId]		The on-floor lineup's players, if/when matched (see above).
score	Score		The score at the time of the shot, team-oriented.
min	float		The ascending game-clock time (minutes) of the shot.
loc	ShotLocation		The shot's transformed court location, in feet.
geo	ShotGeo		The shot's synthetic lat/lon.
dist	float		The shot's distance from the basket, in feet.
pts	int		Made(1)/missed(0) flag from this task -- NOT the real point value (see above).
value	int		Always 0 from this task (see above).
ast_by	Optional[PlayerCodeId]		The assisting player, if/when matched (see above).
is_ast	Optional[bool]		Whether the shot was assisted, if/when matched (see above).
is_trans	Optional[bool]		Whether the shot was in transition, if/when matched (see above).
raw_event	Optional[str]		The raw SVG <title> text this shot was parsed from, for debugging (discarded before writing to disk upstream).

ShotGeo

ShotGeo(lat: 'float', lon: 'float') -> None

A shot's synthetic lat/lon, for geo-aware visualization tooling

(ShotEvent.ShotGeo, models/ncaa/ShotEvent.scala:45). Scope addition, Task 5e.5 -- flattened per ShotLocation's note.

Parameters

Parameter	Type	Default	Description
lat	float		Synthetic latitude (feet-to-meters converted, offset from an arbitrary base point -- not a real-world location).
lon	float		Synthetic longitude, same convention as lat.

ShotLocation

ShotLocation(x: 'float', y: 'float') -> None

A shot's court-relative coordinates, in feet (ShotEvent.ShotLocation,

models/ncaa/ShotEvent.scala:48). Scope addition, Task 5e.5 -- flattened out of the Scala ShotEvent companion object per this module's established nested-object-flattening precedent (see the module docstring's "Scala idiom decisions": ScoreInfo/PlayerCodeId were already flattened out of LineupEvent's companion the same way).

Parameters

Parameter	Type	Default	Description
x	float		Feet from the basket; positive is to the right of the basket (facing the goal), negative is to the left.
y	float		Feet from the basket along the baseline-perpendicular axis.

TeamId

TeamId(name: 'str') -> None

CBB team identifier (TeamId, TeamId.scala, AnyVal).

Parameters

Parameter	Type	Default	Description
name	str		The unique team name.

TeamSeasonId

TeamSeasonId(team: 'TeamId', year: 'Year') -> None

A team's season identifier (TeamSeasonId, TeamSeasonId.scala).

Parameters

Parameter	Type	Default	Description
team	TeamId		The team playing the season.
year	Year		The year the season ends.

adjust_efficiency

adjust_efficiency(game_eff: 'pl.DataFrame', *, league: 'str' = 'mens', max_iter: 'int' = 100, tol: 'float' = 0.0001) -> 'pl.DataFrame'

Iterative opponent-adjusted efficiency -> AdjO / AdjD / AdjEM per team-season.

KenPom-style fixed point: initialise adj_o = raw_o / adj_d = raw_d, then repeatedly recompute each team's rating from its games with the opponent's current adjusted rating and a home-court adjustment removed, until the largest change is below tol. Ratings are computed independently per season (a team's opponent pool is within-season).

The per-game offensive update is off_eff - (adj_d_opp - avg) - loc_o where loc_o is +hfa/2 at home, -hfa/2 away, 0 neutral (defense is symmetric with the opposite sign); avg is the league mean efficiency and hfa comes from ~sportsdataverse.mbb.mbb_prediction_constants.get_constants.

Parameters

Parameter	Type	Default	Description
game_eff	DataFrame		Output of raw_game_efficiency.
league	str	'mens'	"mens" / "womens" -- selects the HFA constant.
max_iter	int	100	Maximum fixed-point iterations.
tol	float	0.0001	Convergence tolerance on the largest rating change.

Returns

One row per (season, team_id): season, team_id, adj_o, adj_d, adj_em, raw_o, raw_d, games. Empty input returns that schema with zero rows.

col_name	type	description
season	integer	Season year.
team_id	character	Unique team identifier.
adj_o	double	Adj o.
adj_d	double	Adj d.
adj_em	double	Adj em.
raw_o	double	Raw o.
raw_d	double	Raw d.
games	integer	Games played.

Example

from sportsdataverse.mbb.mbb_team_ratings import adjust_efficiency, raw_game_efficiency
ratings = adjust_efficiency(raw_game_efficiency(sched, box))

adjust_off_rating_stats

adjust_off_rating_stats(pts_correction_factor: 'float', poss_correction_factor: 'float', mutable_o_rtg: 'ORtgDiagnostics', maybe_raw_o_rtg: 'float | None') -> 'tuple[float, float] | None'

Apply a missing-possession correction factor to an ORtgDiagnostics dict in place.

Faithful port of RatingUtils.adjustOffRatingStats (RatingUtils.ts:993-1033). Genuinely public upstream (called from LineupTableUtils.ts after a lineup-level pts/poss reconciliation), so this port is public too. Recomputes the productivity fields via build_productivity (reused, not re-derived).

Landmine 4 (see module docstring): the o_adj = avgEff / defSos or 1 recomputation here is unguarded against defSos == 0 -- same reachability analysis as landmine 3 (only reachable if the diagnostics dict's original build_o_rtg call used avg_efficiency == 0).

Parameters

Parameter	Type	Default	Description
pts_correction_factor	float		Points correction factor (e.g. team pts / sum of player pts, capped to [0.95, 1.05] by callers).
poss_correction_factor	float		Possession correction factor, same shape.
mutable_o_rtg	ORtgDiagnostics		The ORtgDiagnostics dict to mutate in place (oRtg, Usage, adjORtg, adjORtgPlus, Usage_Bonus, SoS_Bonus, adjPtsFactor, adjPossFactor, and (conditionally) Raw_Usage are all updated).
maybe_raw_o_rtg	float | None		The un-overridden raw oRtg value (rawORtg's .value, or None when no override was in play), used to compute the raw-side return.

Returns

(new_raw_o_rtg, raw_adj_o_rtg_plus) when both mutable_o_rtg["Raw_Usage"] and maybe_raw_o_rtg are not None; otherwise None (.isNilsemantics -- an explicit0` does NOT count as nil).

Example

from sportsdataverse.mbb.mbb_ratings import build_o_rtg, adjust_off_rating_stats

_, _, raw_o_rtg, _, o_diags = build_o_rtg(player, {}, {}, 100.0, True, False)
maybe_raw = raw_o_rtg["value"] if raw_o_rtg else None
adjust_off_rating_stats(1.1, 0.9, o_diags, maybe_raw)
print(o_diags["oRtg"], o_diags["adjORtgPlus"])

adjust_tempo

adjust_tempo(game_eff: 'pl.DataFrame', *, league: 'str' = 'mens', max_iter: 'int' = 100, tol: 'float' = 0.0001) -> 'pl.DataFrame'

Opponent-adjusted tempo (possessions/40) per team-season.

Same fixed point as adjust_efficiency, applied to game possessions under the additive model poss = tempo_i + tempo_j - avg: a team's tempo is recovered by removing its opponents' current adjusted tempo. avg is the league baseline tempo from ~sportsdataverse.mbb.mbb_prediction_constants.get_constants.

Parameters

Parameter	Type	Default	Description
game_eff	DataFrame		Output of raw_game_efficiency.
league	str	'mens'	"mens" / "womens" -- selects the tempo baseline.
max_iter	int	100	Maximum fixed-point iterations.
tol	float	0.0001	Convergence tolerance on the largest tempo change.

Returns

One row per (season, team_id): season, team_id, adj_tempo. Empty input returns that schema with zero rows.

col_name	type	description
season	integer	Season year.
team_id	character	Unique team identifier.
adj_tempo	double

Example

from sportsdataverse.mbb.mbb_team_ratings import adjust_tempo, raw_game_efficiency
tempo = adjust_tempo(raw_game_efficiency(sched, box))

aggregate_player_seasons

aggregate_player_seasons(seasons: "'list[int]'", *, league: 'str' = 'mens') -> 'pl.DataFrame'

Canonical per-player-season counting frame from the boxscore release.

Sums the per-game player boxscores into one row per (player_id, season, team_id) with the counting columns player_per100_features expects. Shot-location splits come from the shots release (2025+): free throws (MadeFreeThrow) are excluded, layup/dunk/tip = rim, and jump shots split three vs mid by score_value (the release's type_text carries no three-point marker; score_value is populated on misses too). For seasons without shots data, three-point attempts come from the box and all remaining attempts fold into fga_mid.

Parameters

Parameter	Type	Default	Description
seasons	list[int]		Seasons to aggregate.
league	str	'mens'	"mens" or "womens".

Returns

One row per (player_id, season, team_id): player_id:Utf8, season, team_id:Utf8, player, minutes + the counting columns + fga_rim, fga_mid, fga_three. Empty input returns zero rows.

col_name	type	description
season	integer	Season year.
team_id	character	Unique team identifier.
player	character	Player name.
position	character	Listed roster position (G, F, C, etc.).
minutes	double	Minutes played, formatted MM:SS (V3 PT-duration parsed) or decimal minutes (V2).
field_goals_made	double	Field goals made (2-pt + 3-pt).
field_goals_attempted	double	Field goal attempts (2-pt + 3-pt).
three_point_field_goals_made	double	Three-point field goals made.
three_point_field_goals_attempted	double	Three-point field goal attempts.
free_throws_made	double	Free throws made.
free_throws_attempted	double	Free throw attempts.
offensive_rebounds	double	Offensive rebounds.
defensive_rebounds	double	Defensive rebounds.
assists	double	Total assists.
steals	double	Total steals.
blocks	double	Total blocks.
turnovers	double	Total turnovers.
points	double	Points scored.
player_id	character	Unique player identifier.
fga_rim	double
fga_mid	double
fga_three	double

Example

from sportsdataverse.mbb.mbb_player_value_constants import (
    aggregate_player_seasons, player_per100_features,
)
feats = player_per100_features(aggregate_player_seasons([2025]))

apply_weak_priors

apply_weak_priors(field: 'str', player_poss_pcts: 'list[float]', prior_info: 'RapmPriorInfo', debug_mode: 'bool' = False) -> 'Callable[[float, list[float]], list[float]]'

Build a closure that nudges ridge-regressed RAPM back towards its weak prior.

Faithful port of RapmUtils.applyWeakPriors (RapmUtils.ts:921-995). Ridge regression depresses estimates towards 0; this "fills" the team-total error (see pick_ridge_regression's [IMPORTANT-EQUATION-01] team-total reconciliation) back in using each player's weak (KenPom-derived) prior as the fallback signal, capped so no more than half the team-total error gets attributed via this path (max_multiplier = -0.5) -- an alternate flat-translation path (use_alt_rating) kicks in for off_adj_ppp/def_adj_ppp fields when the capped path can't fully explain the error.

Parameters

Parameter	Type	Default	Description
field	str		The prior key to read off each prior_info["players_weak"] entry, e.g. "off_adj_ppp".
player_poss_pcts	list[float]		Per-player possession-share weights (index-aligned with prior_info["players_weak"]), e.g. pick_ridge_regression's own pct_by_player[off_or_def].
prior_info	RapmPriorInfo		A RapmPriorInfo (only ["players_weak"] is read).
debug_mode	bool	False	Kept for TS signature parity -- upstream gates a console.log behind this flag (RapmUtils.ts:979-984), which this port deliberately does not reproduce: every production call site pins it False (offDefDebugMode.off/.def are hardcoded False constants inside pickRidgeRegression), so it is dead in every current caller and would only ever emit console noise, not test-observable behavior.

Returns

A closure (error, base_results) -> adjusted_results -- call it with the team-total efficiency error and the pre-adjustment RAPM vector to get the weak-prior-nudged result.

Example

from sportsdataverse.mbb.mbb_rapm import apply_weak_priors

nudge = apply_weak_priors("off_adj_ppp", pct_by_player, ctx["prior_info"])
adjusted = nudge(adj_eff_err_pre_prior, results_pre_prior)

as_of_ratings_split

as_of_ratings_split(results: 'pl.DataFrame', cutoff_date: 'datetime.date') -> 'pl.DataFrame'

Filter a results frame to games strictly before a cutoff date (leakage boundary).

Parameters

Parameter	Type	Default	Description
results	DataFrame		A polars.DataFrame with a date column.
cutoff_date	date		Games on or after this date are excluded.

Returns

A polars.DataFrame containing only rows with date < cutoff_date.

col_name	type	description
id	integer	Id.
uid	character	ESPN UID string.
attendance	double	Reported attendance.
time_valid	logical	Time valid.
neutral_site	logical	Neutral site.
conference_competition	logical	Conference competition.
play_by_play_available	logical
recent	logical	Recent.
start_date	character	Start date (YYYY-MM-DD).
notes_type	character	Notes type.
notes_headline	character	Notes headline.
broadcast_market	character	Broadcast market label (e.g. 'national', 'home').
broadcast_name	character	Broadcast name.
type_id	integer	Type identifier (numeric).
type_abbreviation	character	Type abbreviation.
venue_id	integer	Unique venue identifier.
venue_full_name	character	Venue full name.
venue_address_city	character	Venue address city.
venue_address_state	character	Venue address state / region.
venue_indoor	logical	TRUE if the venue is indoors.
status_clock	double	Status clock.
status_display_clock	character	Status display clock.
status_period	double	Status period.
status_type_id	integer	Unique identifier for status type.
status_type_name	character	Status type name.
status_type_state	character	Status type state.
status_type_completed	logical	Status type completed.
status_type_description	character	Status type description.
status_type_detail	character	Status type detail.
status_type_short_detail	character	Status type short detail.
format_regulation_periods	double	Format regulation periods.
home_team_id	integer	Unique identifier for the home team.
home_uid	character	Home team's uid.
home_location	character	Home team's location.
home_name	character	Home name.
home_abbreviation	character	Home team's abbreviation.
home_display_name	character	Home display name.
home_short_display_name	character	Home short display name.
home_color	character	Color code (hex) for home.
home_alternate_color	character	Color code (hex) for home alternate.
home_is_active	logical	Home team's is active.
home_venue_id	integer	Unique identifier for home venue.
home_logo	character	Home team logo URL.
home_conference_id	integer	Unique identifier for home conference.
home_score	integer	Home team score at the time of the play.
home_winner	logical	Home team's winner.
home_current_rank	double
home_linescores	character
home_records	character
away_team_id	integer	Unique identifier for the away team.
away_uid	character	Away team's uid.
away_location	character	Away team's location.
away_name	character	Away name.
away_abbreviation	character	Away team's abbreviation.
away_display_name	character	Away display name.
away_short_display_name	character	Away short display name.
away_color	character	Color code (hex) for away.
away_alternate_color	character	Color code (hex) for away alternate.
away_is_active	logical	Away team's is active.
away_venue_id	integer	Unique identifier for away venue.
away_logo	character	Away team logo URL.
away_conference_id	integer	Unique identifier for away conference.
away_score	integer	Away team score at the time of the play.
away_winner	logical	Away team's winner.
away_current_rank	double
away_linescores	character
away_records	character
game_id	integer	Unique game identifier.
season	integer	Season year.
season_type	integer	Season type (1=pre-season, 2=regular season, 3=postseason, 4=off-season for ESPN; or string label for WNBA Stats).
status_type_alt_detail	character	Status type alt detail.
tournament_id	integer	ESPN tournament identifier.
groups_id	integer	Unique identifier for groups.
groups_name	character	Groups name.
groups_short_name	character	Groups short name.
groups_is_conference	logical	Groups is conference.
game_json	logical
game_json_url	character
game_date_time	character	Game start date/time (ISO 8601).
date	character	Date in YYYY-MM-DD format.
PBP	logical
team_box	logical	Team box.
player_box	logical	Player box.

Example

import datetime as dt
from sportsdataverse._common.metrics import as_of_ratings_split
as_of_ratings_split(results, dt.date(2023, 9, 8))

as_of_season_split

as_of_season_split(df: 'pl.DataFrame', target_season: 'int') -> 'pl.DataFrame'

Rows strictly before target_season -- the leakage boundary.

Parameters

Parameter	Type	Default	Description
df	DataFrame		Frame with an integer season column.
target_season	int		The season being predicted; its rows (and later) drop.

Returns

The subset with season < target_season.

col_name	type	description
season	integer	Season year.
team_id	character	Unique team identifier.
player	character	Player name.
position	character	Listed roster position (G, F, C, etc.).
minutes	double	Minutes played, formatted MM:SS (V3 PT-duration parsed) or decimal minutes (V2).
field_goals_made	double	Field goals made (2-pt + 3-pt).
field_goals_attempted	double	Field goal attempts (2-pt + 3-pt).
three_point_field_goals_made	double	Three-point field goals made.
three_point_field_goals_attempted	double	Three-point field goal attempts.
free_throws_made	double	Free throws made.
free_throws_attempted	double	Free throw attempts.
offensive_rebounds	double	Offensive rebounds.
defensive_rebounds	double	Defensive rebounds.
assists	double	Total assists.
steals	double	Total steals.
blocks	double	Total blocks.
turnovers	double	Total turnovers.
points	double	Points scored.
player_id	character	Unique player identifier.
fga_rim	double
fga_mid	double
fga_three	double

Example

from sportsdataverse.mbb.mbb_player_value_constants import as_of_season_split
prior = as_of_season_split(df, 2026)

bootstrap_ari

bootstrap_ari(fit_fn: "'Callable[[np.ndarray], tuple[np.ndarray, np.ndarray]]'", X: 'np.ndarray', n_boot: 'int' = 20, seed: 'int' = 0) -> 'float'

Cluster stability: mean ARI between the full fit and bootstrap refits.

Parameters

Parameter	Type	Default	Description
fit_fn	Callable[[ndarray], tuple[ndarray, ndarray]]		X -> (centers, labels) (e.g. a seeded kmeans_fit partial).
X	ndarray		Feature matrix.
n_boot	int	20	Bootstrap resamples.
seed	int	0	RNG seed.

Returns

Mean adjusted Rand index of the resample fits' assignments (of the FULL sample, via nearest refit center) vs the full-fit labels.

Example

from functools import partial
score = bootstrap_ari(lambda Z: kmeans_fit(Z, 8, seed=0), Z, n_boot=20, seed=0)

brier_score

brier_score(y_true: 'np.ndarray', p_pred: 'np.ndarray') -> 'float'

Mean squared error between predicted probabilities and binary outcomes.

Parameters

Parameter	Type	Default	Description
y_true	ndarray		Array of binary outcomes (0/1).
p_pred	ndarray		Array of predicted probabilities in [0, 1].

Returns

The Brier score (0.0 is a perfect forecast).

Example

import numpy as np
from sportsdataverse._common.metrics import brier_score
brier_score(np.array([1, 0]), np.array([0.9, 0.1]))

build_d_rtg

build_d_rtg(stat_set: 'LineupStatSet | None', avg_efficiency: 'float', calc_diags: 'bool', override_adjusted: 'bool') -> 'tuple[dict[str, float] | None, dict[str, float] | None, dict[str, float] | None, dict[str, float] | None, DRtgDiagnostics | None]'

Individual defensive rating (Dean-Oliver DRtg) + diagnostics.

Faithful port of RatingUtils.buildDRtg (RatingUtils.ts:1252-1485). Mirrors build_o_rtg's structure (stat_get closure, calc_diags/override_adjusted flag pair, recursive un-overridden raw-value pass) over the simpler (stat_set, avg_efficiency, calc_diags, override_adjusted) 4-arg signature (no roster/extra-team-stat args, confirmed against the TS).

Parameters

Parameter	Type	Default	Description
stat_set	LineupStatSet | None		The player's stat dict. None returns an all-None 5-tuple (RatingUtils.ts:1264-1265's if (!statSet) -- null/undefined only). Unlike build_o_rtg, an empty dict computes cleanly -- every division in buildDRtg is guard-ternary'd (see the module docstring's "Contrast" note), so {} does not raise ZeroDivisionError.
avg_efficiency	float		League/context average efficiency (100 in every vendored jest call).
calc_diags	bool		When True, populate the 5th tuple slot (DRtgDiagnostics); otherwise it is None.
override_adjusted	bool		When True, apply build_def_overridesto the raw opponent-FGM/points fields before computing, and additionally recurse once (withcalc_diags=False, override_adjusted=False`) to compute the un-overridden "raw" values for the 3rd/4th tuple slots.

Returns

A 5-tuple (d_rtg, adj_d_rtg, raw_d_rtg, raw_adj_d_rtg, d_rtg_diags): - d_rtg: {"value": DRtg} when Opponent_Possessions_Box > 0, else None. - adj_d_rtg: {"value": Adj_DRtgPlus} under the same guard. - raw_d_rtg / raw_adj_d_rtg: the un-overridden values from the recursive call when override_adjusted=True; None otherwise (unlike build_o_rtg, there is no internal-usage special case here -- the TS destructures only the first 2 slots of the recursive 5-tuple). - d_rtg_diags: the full DRtgDiagnostics dict (None unless calc_diags=True).

Example

from sportsdataverse.mbb.mbb_ratings import build_d_rtg

d_rtg, adj_d_rtg, _, _, diags = build_d_rtg(player, 100.0, True, False)
print(d_rtg["value"], diags["dRtg"])

# Override-adjusted (manual 3P-defense-% override applied)

d_rtg2, adj_d_rtg2, raw_d_rtg2, raw_adj_d_rtg2, _ = build_d_rtg(
    player, 100.0, False, True,
)

build_mbb_season_wp

build_mbb_season_wp(season: 'int', *, league: 'str' = 'mens', return_as_pandas: 'bool' = False) -> "Union[pl.DataFrame, 'pd.DataFrame']"

A season's play-by-play with win-probability columns joined in.

Loads the season's play-by-play, schedule, and team boxscores, builds a leakage-free weekly as-of pregame anchor per game, scores every play through the bundled in-game win-probability artifact, and returns the full load_mbb_pbp frame with pregame_home_prob + home_win_prob appended -- the enrich-in-place shape that overwrites the season's play_by_play_<season>.parquet release asset.

Parameters

Parameter	Type	Default	Description
season	int		Season year (e.g. 2024); bounded by load_mbb_pbp release availability (>= 2002).
league	str	'mens'	"mens" or "womens" (selects the loaders + constants).
return_as_pandas	bool	False	Return a pandas DataFrame instead of polars.

Returns

The season's load_mbb_pbp frame (every column preserved) with the two WP_COLSappended (bothFloat64), sorted by game_idthengame_play_number`.

col_name	type	description
game_play_number	integer	Sequential play number within the game.
id	integer	Id.
sequence_number	integer	Sequence number representing a shot-possession (V3 PBP).
type_id	integer	Type identifier (numeric).
type_text	character	Display text for the type field.
text	character	Text description of the play / record.
away_score	integer	Away team score at the time of the play.
home_score	integer	Home team score at the time of the play.
period_number	integer	Numeric period (1-4 for quarters; 5+ for OT).
period_display_value	character	Period display label (e.g. '1st Quarter', 'OT').
clock_display_value	character	Game clock display string (e.g. '8:32').
scoring_play	logical	TRUE if the play resulted in points scored.
score_value	integer	Point value of the play (2 / 3 / 1).
team_id	integer	Unique team identifier.
athlete_id_1	integer	Primary athlete identifier (e.g. shooter).
wallclock	character	Wallclock.
shooting_play	logical	TRUE if the play was a shooting attempt.
athlete_id_2	integer	Secondary athlete identifier (e.g. assister / fouler).
game_id	integer	Unique game identifier.
season	integer	Season year.
season_type	integer	Season type (1=pre-season, 2=regular season, 3=postseason, 4=off-season for ESPN; or string label for WNBA Stats).
home_team_id	integer	Unique identifier for the home team.
home_team_name	character	Home team name.
home_team_mascot	character	Home team mascot.
home_team_abbrev	character	Home team three-letter abbreviation.
home_team_name_alt	character	Alternate home team name.
away_team_id	integer	Unique identifier for the away team.
away_team_name	character	Away team name.
away_team_mascot	character	Away team mascot.
away_team_abbrev	character	Away team three-letter abbreviation.
away_team_name_alt	character	Alternate away team name.
game_spread	double	Game spread (signed; positive = home favored).
home_favorite	logical	TRUE if the home team is the betting favorite.
game_spread_available	logical	TRUE if a point spread was available.
home_team_spread	double	Home team's point spread.
half	integer	Half of the game (1 or 2).
time	character	Time / clock value.
clock_minutes	integer	Clock minutes split out for convenience.
clock_seconds	integer	Clock seconds split out for convenience.
home_timeout_called	logical
away_timeout_called	logical
lag_period	integer
lead_period	integer
lag_half	integer	A lag column on the half
lead_half	integer	A lead column on the half
start_period_seconds_remaining	integer
start_game_seconds_remaining	integer	Seconds remaining in the game at the start of the play.
end_period_seconds_remaining	integer
end_game_seconds_remaining	integer	Seconds remaining in the game at the end of the play.
game_date	character	Game date (YYYY-MM-DD).
game_date_time	character	Game start date/time (ISO 8601).
coordinate_x	double	X coordinate on the court (half-court layout).
coordinate_y	double	Y coordinate on the court (half-court layout).
coordinate_x_raw	double	X coordinate as returned by the API before any adjustment.
coordinate_y_raw	double	Y coordinate as returned by the API before any adjustment.
athlete_name_1	character
athlete_name_2	character
athlete_name_3	character
media_id	character	Media identifier (video / image).
pregame_home_prob	double
home_win_prob	double

Example

from sportsdataverse.mbb import build_mbb_season_wp
wp = build_mbb_season_wp(2024)
wp.select("game_id", "game_play_number", "home_win_prob").head()

build_net_points

build_net_points(player_rapm_and_poss_pct: 'LineupStatSet', ortg: 'ORtgDiagnostics', drtg: 'DRtgDiagnostics', avg_eff: 'float', scale_type: "Literal['T%', 'P%', '/G']", num_games: 'float' = 1, missing_game_adjustment: 'float' = 1) -> 'NetPoints'

Decompose ORtg/DRtg + RAPM into a Net-Points-like breakdown.

Faithful port of RatingUtils.buildNetPoints (RatingUtils.ts:1036-1234). Genuinely public upstream (called from buildLeaderboards.ts, PlayerImpactBreakdownTable.tsx, and ImpactBreakdownUtils.ts), so this port is public too.

Parameters

Parameter	Type	Default	Description
player_rapm_and_poss_pct	LineupStatSet		The player's stat dict -- reads off_team_poss_pct/def_team_poss_pct (nullish-coalesced to 0.0, see nullish) and, when present, off_adj_rapm/def_adj_rapm(each a{"value": float}` "Statistic"-shaped field) for the RAPM "with-or-without-you" (WOWY) deltas.
ortg	ORtgDiagnostics		An ORtgDiagnostics dict from build_o_rtg (calc_diags=True), typically with adjPtsFactor/ adjPossFactor overridden from their 1 default by a missing-possession correction.
drtg	DRtgDiagnostics		A DRtgDiagnostics dict from build_d_rtg (calc_diags=True). If it carries an onBallDiags key (this port's build_d_rtg never sets one -- see the module docstring's deferred-work note), the on-ball-adjusted branch is used instead of the base dRtg/adjDRtgPlus.
avg_eff	float		League/context average efficiency.
scale_type	Literal['T%', 'P%', '/G']		"T%" (scale by on-floor team-possession share, avgEff-adjusted possession count), "P%" (scale to 100 possessions), or "/G" (scale to per-game).
num_games	float	1	Divisor for the "/G" scale type. Default 1.
missing_game_adjustment	float	1	Multiplier folded into the "T%" scale factor for imputed-missing-games correction. Default 1.

Returns

A NetPoints dict -- 20 keys, plus an optional defNetPtsIndiv 21st key present only when drtg["onBallDiags"] is set (TS-verbatim key names throughout).

Example

from sportsdataverse.mbb.mbb_ratings import build_o_rtg, build_d_rtg, build_net_points

_, _, _, _, o_diags = build_o_rtg(player, {}, {}, 100.0, True, False)
_, _, _, _, d_diags = build_d_rtg(player, 100.0, True, False)
net_pts = build_net_points(player, o_diags, d_diags, 100.0, "T%")
print(net_pts["offNetPts"], net_pts["defNetPts"])

build_o_rtg

build_o_rtg(stat_set: 'LineupStatSet | None', roster_stats_by_code: 'dict[str, LineupStatSet] | None', extra_team_stat_info: 'LineupStatSet', avg_efficiency: 'float', calc_diags: 'bool', override_adjusted: 'bool') -> 'tuple[dict[str, float] | None, dict[str, float] | None, dict[str, float] | None, dict[str, float] | None, ORtgDiagnostics | None]'

Individual offensive rating (Dean-Oliver ORtg) + diagnostics.

Faithful port of RatingUtils.buildORtg (RatingUtils.ts:398-960). See the module docstring for the signature-vs-brief note (this mirrors the TS 6-positional-arg / 5-tuple contract verbatim, snake_cased) and the diagnostics-dict key-naming convention (TS-verbatim, not snake_cased).

Parameters

Parameter	Type	Default	Description
stat_set	LineupStatSet | None		The player's LineupStatSet (ES-aggregation-shaped per-player doc, "IndivStatSet" upstream). None returns an all-None 5-tuple (RatingUtils.ts:412-413's if (!statSet) -- null/undefined only). An empty dict does NOT short-circuit ({} is truthy in JS and falls through to compute upstream); in this port it falls into unguarded-division landmine 1 and raises ZeroDivisionError where the TS degrades to a NaN-laced degenerate result -- see the module docstring's landmine list.
roster_stats_by_code	dict[str, LineupStatSet] | None		{player_code: LineupStatSet} for every player on the roster -- used for the approximate team-ORB apportionment and the per-shot-location assisted-eFG fallback. None is treated as {} (every vendored jest call passes a literal {}).
extra_team_stat_info	LineupStatSet		{"total_off_to": {...}, "sum_total_off_to": {...}} -- team-level TOV bookkeeping used to compute "unblamed" team turnovers apportioned by off_team_poss_pct.
avg_efficiency	float		League/context average efficiency (100 in every vendored jest call).
calc_diags	bool		When True, populate the 5th tuple slot (ORtgDiagnostics); otherwise it is None.
override_adjusted	bool		When True, apply build_off_overridesto the raw made/attempt/turnover fields before computing, and additionally recurse once (withcalc_diags=False, override_adjusted=False`) to compute the un-overridden "raw" values for the 3rd/4th tuple slots.

Returns

A 5-tuple (o_rtg, adj_o_rtg, raw_o_rtg, raw_adj_o_rtg, o_rtg_diags): - o_rtg: {"value": ORtg} when TotPoss > 0, else None. - adj_o_rtg: {"value": Adj_ORtgPlus} when TotPoss > 0, else None. - raw_o_rtg: when calc_diags or override_adjusted, the un-overridden ORtg (None if override_adjusted=False, since no un-overridden pass was computed); otherwise a special internal-recursion value {"value": usage} (RatingUtils.ts:835's "if called internally return usage here" case). - raw_adj_o_rtg: the un-overridden adj_o_rtg (None when override_adjusted=False). - o_rtg_diags: the full ORtgDiagnostics dict (None unless calc_diags=True).

Example

from sportsdataverse.mbb.mbb_ratings import build_o_rtg

o_rtg, adj_o_rtg, _, _, diags = build_o_rtg(
    player, {}, {"total_off_to": {"value": 0}, "sum_total_off_to": {}},
    100.0, True, False,
)
print(o_rtg["value"], diags["oRtg"])

# Override-adjusted (manual shooting-% overrides applied)

o_rtg2, adj_o_rtg2, raw_o_rtg2, raw_adj_o_rtg2, _ = build_o_rtg(
    player, {}, {"total_off_to": {"value": 0}, "sum_total_off_to": {}},
    100.0, False, True,
)

build_player_context

build_player_context(players: 'list[PlayerOnOffStats]', lineups: 'list[LineupStatSet]', players_baseline: 'dict[PlayerId, IndivStatSet]', stats_averages: 'PureStatSet', avg_efficiency: 'float', agg_value_key: 'ValueKey' = 'value', config: 'RapmConfig' = {'prior_mode': -1, 'removal_pct': 0.06, 'fixed_regression': -1}) -> 'RapmPlayerContext'

Build the context object the RAPM matrix-solve layer consumes.

Faithful port of RapmUtils.buildPlayerContext (RapmUtils.ts:427-541). Removes low-possession players (config["removal_pct"] of total on+off possessions), flags fully-removed lineups (mutating lineups in place -- see the module docstring's landmine 5), builds the player-to-column index, and folds build_priors into prior_info.

Parameters

Parameter	Type	Default	Description
players	list[PlayerOnOffStats]		The per-player on/off splits (PlayerOnOffStats), e.g. mbb_lineup_stats.lineup_to_team_report(...)["players"].
lineups	list[LineupStatSet]		The per-lineup LineupStatSet docs feeding this team's aggregate (mutated in place -- see landmine 5).
players_baseline	dict[PlayerId, IndivStatSet]		{player_id: IndivStatSet} -- forwarded to build_priors unchanged.
stats_averages	PureStatSet		League/context average stat set -- forwarded to build_priors unchanged.
avg_efficiency	float		League/context average efficiency.
agg_value_key	ValueKey	'value'	"value" or "old_value" -- forwarded to build_priors as its value_key (only affects prior calculations, not the lineup-filtering/aggregation above it).
config	RapmConfig	{'prior_mode': -1, 'removal_pct': 0.06, 'fixed_regression': -1}	Removal-percent / prior-mode / regression config. Defaults to DEFAULT_RAPM_CONFIG; never mutated by this function (only config["removal_pct"]/config["prior_mode"] are read), matching the TS default parameter's own read-only usage.

Returns

A RapmPlayerContext.

Example

from sportsdataverse.mbb.mbb_lineup_stats import lineup_to_team_report
from sportsdataverse.mbb.mbb_rapm import build_player_context, DEFAULT_RAPM_CONFIG

report = lineup_to_team_report({"lineups": buckets, "error_code": None})
ctx = build_player_context(
    report["players"], buckets, {}, {}, 100.0, "value", DEFAULT_RAPM_CONFIG
)
print(ctx["num_players"], ctx["team_info"]["off_poss"]["value"])

# Filtering lineups by side (the ``filtered_lineups`` closure)

off_lineups = ctx["filtered_lineups"]("off")
def_lineups = ctx["filtered_lineups"]("def")

build_priors

build_priors(players_baseline: 'dict[PlayerId, IndivStatSet]', stats_averages: 'PureStatSet', avg_efficiency: 'float', col_to_player: 'list[str]', prior_mode: 'float', value_key: 'ValueKey' = 'value') -> 'RapmPriorInfo'

Build strong/weak per-player RAPM priors for every column.

Faithful port of RapmUtils.buildPriors (RapmUtils.ts:237-407). See the module docstring's landmine list, item 1, for the critical Python-vs-JS {}-truthiness gotcha this function's implementation deliberately avoids.

Parameters

Parameter	Type	Default	Description
players_baseline	dict[PlayerId, IndivStatSet]		{player_id: IndivStatSet} -- the most-general per-player baseline info (in production, sourced from mbb_ratings.build_productivity's output; see the module docstring's "RAPM prior source" note).
stats_averages	PureStatSet		League/context average stat set, used by the (currently dead-code, see landmine 4) get_prior_basis fallback and by with_avg_or_undef's nil-check gate.
avg_efficiency	float		League/context average efficiency.
col_to_player	list[str]		The player ids, in column order -- playersStrong/ playersWeak are index-aligned with this list.
prior_mode	float		-1 for adaptive mode, -2 (or lower) for no prior, 0-1 for a fixed strong-prior weight.
value_key	ValueKey	'value'	"value" or "old_value" -- allows priors to be built from luck-adjusted parameters.

Returns

A RapmPriorInfo.

Example

from sportsdataverse.mbb.mbb_rapm import build_priors

priors = build_priors({}, {}, 100.0, ["Wiggins, Aaron"], -1)
print(priors["players_weak"][0])