# Provenance, coverage and caveats

**Attribution:** Virtually Varsity (virtuallyvarsity.com), CC BY 4.0

## What is licensed, and what is not

The CC BY 4.0 licence covers Virtually Varsity's **compilation** of the data: the
scraping, cleaning, de-duplication, identity resolution and canonicalisation.
**No ownership is claimed over the underlying facts.** Game scores, statistics, standings,
poll ranks and honors are facts of public record, published by the leagues and sites listed
below. Team names, marks and logos (`teams.logo_url`) belong to their owners and are not
licensed by this release. Virtually Varsity is an independent archive and is not affiliated
with or endorsed by the MCLA or any team.

## Sources

| Source | What it contributed | Tables |
|---|---|---|
| **mcla.us** (Men's Collegiate Lacrosse Association, live site) | Season schedules and scores (2006+), conference membership, standings, weekly polls (2009+), per-game detail pages (box score, team stats, player stats), player pages and season totals (2015+), live All-America pages (2025+) | `games`, `team_seasons`, `polls`, `game_box_score`, `game_team_stats`, `game_player_stats`, `players`, `player_seasons`, `player_awards` (`source='mcla-us'`), `teams` |
| **archive.mcla.us** (mirror of the former MCLA site) | Player season totals, per-game player lines and honors for 2006–2024. These are more complete than the live site, so they **take precedence** for the years they cover. | `player_seasons`, `game_player_stats`, `players` (`arc-…` ids), `player_awards` (`source='archive-mcla'`) |
| **uslia.net** (US Lacrosse Intercollegiate Associates, the MCLA's predecessor) | Legacy game scores 1997–2005 and Top-25 polls 2000–2006 (rank + team name only) | `games` (`uslia-…` ids), `polls` |
| **laxpower.com**, via the Internet Archive Wayback Machine | All-American teams 2008–2017 (site defunct ~2016). These rows are mostly superseded by the sources above, and only rows with no better source remain. | `player_awards` (`source='laxpower-wayback'`) |
| **MCLA Division I All-America list** (CSV compiled from MCLA records) | Authoritative D-I All-Americans 2000–2023, incl. honorable mention | `player_awards` (`source='allamerica-csv'`) |
| **mcla.us poll pages via Wayback** / **CollegeLAX.us via Wayback** | 2014–2015 polls (no longer served live) / 2007–2008 MCLA Top-25 news posts | `polls` |
| **MCLA national tournament bracket history** (CSV) | Brackets with seeds, 1997–present | `tournament_games`, and `games` (`game_type='NT'`, `tg_…` ids for games missing from the log) |
| **Wikipedia** conference articles (RMLC, SELC, WCLL, LSA, …) | Conference championship winners | `team_awards` (`source='wikipedia'`) |
| **OpenStreetMap Nominatim** | Campus coordinates. Location data © OpenStreetMap contributors, ODbL ([openstreetmap.org/copyright](https://www.openstreetmap.org/copyright)). | `teams.latitude/longitude` |
| Derived by Virtually Varsity | National champion / runner-up chips from bracket finals (`team_awards.source='auto'`). Season W-L-T repaired against the game log where standings pages were wrong. | `team_awards`, `team_seasons` |

## Coverage

| Table | Years | Notes |
|---|---|---|
| `games` | 1997–present | 1997–2005 from uslia.net. About 290 games (bracket backfills) have no source URL. |
| `team_seasons` | 1997–present | Conference/division labels are blank on most pre-~2005 rows. |
| `game_box_score` | 2006–present | ~90–100% of played games per year |
| `game_team_stats` | 2015–present | MCLA did not publish team game stats before 2015 |
| `game_player_stats` | 2006–present | 2006–2014 lines come from archive.mcla.us and have no `position_group` / `eligibility` |
| `player_seasons` | 2006–present | Pre-2015 `gp` is a floor. 2006 is the earliest player data that exists anywhere. |
| `players` | careers from 2006 | plus synthetic `aa-…` records for pre-2006 honorees (awards only, no stats) |
| `player_awards` | 2000–present | All-American 2000+, All-Conference from archive.mcla.us |
| `polls` | D1 2000+, D2 2004+ | No online polls exist for some early weeks/seasons |
| `tournament_games` | D1 1997+, D2 2005+ | |
| `team_awards` | conference titles back to 1976, national titles 1997+ | |

## How identifiers are canonicalised

The source sites rename teams and re-issue players under new ids. The release fixes this once
at build time, so a plain `JOIN` gives correct results:

1. **Players.** mcla.us re-mints the same real person under many `<name>-<hash>` ids, often
   one per season. Duplicate identities were merged in two passes. First, byte-identical
   careers were merged. Second, a reviewed set of same-name, same-team, non-overlapping-season
   splits was merged. Archive-only players were linked to their mcla.us identity where one
   exists. Every `player_id` in every fact table is the **canonical** id. The retired ids are
   listed in `player_id_crosswalk`.
2. **Teams.** Renamed or migrated slugs (including mcla.us's 2025 site-wide slug change) and
   same-school duplicate records map to one canonical slug through an exact-match alias table.
   Every `team_slug` / `home_slug` / `away_slug` / `winner_slug` / `loser_slug` is canonical,
   and the retired slugs are listed in `team_slug_crosswalk`.
3. **Player-to-team attribution in box scores.** mcla.us roster tabs use a different slug format
   from its team pages: `brigham_young` vs `brigham-young`, full-name-plus-hash forms, and
   abbreviations like `lsu`. Each player line is therefore attributed to a side in layers:
   - inherit the box score's recorded home/away side;
   - then a normalized slug match;
   - then elimination against the already-attributed team;
   - then a curated alias list plus token similarity.

   **Ties are never guessed.** `game_player_stats.team_slug` is the game's canonical home or
   away slug, and the raw value is kept in `source_team_slug`. This is the same resolution the
   virtuallyvarsity.com game pages use.
4. **Honors** are de-duplicated across sources, spellings and levels, so one honor counts once.
   Each honor is linked to a canonical player where the name, team and year match.

## Known gaps and caveats

- **Box-score periods vs total.** About 2,350 older box-score rows (mostly 2006–08) have
  quarters that don't sum to `total`. This is how MCLA entered the data. `total` is
  authoritative.
- **Pure-bench lines dropped.** `game_player_stats` omits all-zero stat lines, so it is not
  a full roster or games-played record. Use `player_seasons.gp`.
- **Pre-2015 games played** (`player_seasons.gp`) is a floor, derived from recorded game
  lines.
- **Face-off losses** are largely untracked before ~2013, and **goals against** before 2006.
  Some archive seasons logged saves or face-off wins but not the opposing counter, so check
  the denominator before computing a rate stat.
- **Unattributed player lines.** In 4 games, a few dozen player lines could not be placed on
  a side unambiguously. They have `team_slug` and `is_home` set to NULL.
- **Truncated season slugs.** About 200 `player_seasons` rows (2015–2025) still carry a
  truncated upstream team slug (`st`, `southern`, `illinois`, `marymount`, …) that matches
  no `teams.slug`. These rows had no same-season game lines to resolve them. Where a
  truncated row was a stat-for-stat subset of the same player's fuller season row on the
  resolved team, it was dropped as a partial duplicate.
- **Legacy games without detail pages.** About 5,350 pre-MCLA games link to uslia.net and
  have no box score. Six 2017 Cal State Channel Islands game pages returned HTTP 500 upstream
  and have no details.
- **Upstream data-entry artifacts** are kept as published. Examples: eligibility `0th`,
  `game_type` `N/A`, mixed poll `week_key` styles (`3` vs `week-3`), and a few poll
  `team_slug`s with no `teams` row.
- **Unlinked honors.** Some `player_awards` rows have no `player_id` (honoree not matched to a
  player) or no `team_slug` (school not resolved). `player_name` and `school_name` always
  carry the source text.
- **Not included:** user-generated content, site analytics, admin data and the public
  team-correction submissions (personal data).

## Corrections

Spotted an error? Use the "Need updates for your team?" form on any team page at
virtuallyvarsity.com. Corrections flow into the next release.