INTELLIGENCE
How the numbers are made.
Player DNA, Course DNA and Course Fit are descriptive, versioned and built only from observed results.
PLAYER DNA
golf-results-dna/1.0.0
| Dimension | Unit | Definition |
|---|---|---|
| Scoring vs field | strokes per round better than the round field average | Mean of (round field average − player round score) across rounds in full-field leaderboards. Positive = better than the field. Equivalent in construction to round-level total strokes gained; no per-shot categories. |
| Consistency | standard deviation of field-adjusted round scores | Sample standard deviation of the player’s field-adjusted round results. Lower = more consistent. |
| Under-par rounds | % of rounds | Share of observed full-field rounds completed under the published par. |
| Cuts made | % of starts | Starts in full-field leaderboards where the player completed all scheduled rounds. |
| Top-10 rate | % of starts | Starts in full-field leaderboards finishing tenth or better (ties included). |
| Contention | % of starts | Starts in full-field leaderboards finishing fifth or better (ties included). |
| Recent form | strokes per round vs field, last 10 full-field starts | Scoring vs field restricted to the player’s ten most recent full-field starts before the as-of date. |
| Major performance | strokes per round vs field in majors | Scoring vs field restricted to major championships with full-field leaderboards. |
Confidence tiers: rounds — High ≥ 40, Medium ≥ 20, Limited ≥ 8; starts — High ≥ 12, Medium ≥ 6, Limited ≥ 3. Percentiles are mid-rank within one division and window, withheld below Limited or when fewer than 10 players qualify. Windows: last 24 months and all observed since 2000. Only results that ended before the as-of date are used.
13 dimensions not measured yet
- Driving power — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Driving control — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Off-tee value — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Approach — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Iron precision — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Greens in regulation — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Short game — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Scrambling — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Sand play — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Putting — No approved shot-level or statistics source. PGA TOUR statistics are HOLD (access barrier); Data Golf requires a paid licence decision.
- Par-3 performance — Hole-by-hole scores exist only for final-round contenders in approved sources; a full-field hole sample is required.
- Par-4 performance — Hole-by-hole scores exist only for final-round contenders in approved sources; a full-field hole sample is required.
- Par-5 performance — Hole-by-hole scores exist only for final-round contenders in approved sources; a full-field hole sample is required.
COURSE DNA
golf-course-dna/1.0.0
Scoring difficulty, spread and under-par rate come from every round in full-field editions at the venue, compared with other courses in the same division. Winning score and cut line use published values. Each edition’s published setup is a separate layout version.
COURSE FIT
golf-course-fit-descriptive/1.0.0
Descriptive components only: course history, performance on comparably difficult setups, performance on comparable lengths and recent form. Each shows the player value, the course signal and both samples. There is no composite score and no win probability.
SOURCES
Where the data comes from
- Wikidata (CC0 1.0): identities, championship editions, winners, venues
- Wikipedia (CC BY-SA 4.0): leaderboards, round scores, fields, cut lines, course setups, tour schedules
- Wikimedia Commons (per image (CC0 / CC BY / CC BY-SA)): player and course photographs with per-image attribution
Not used: PGA TOUR (access barrier), OWGR (terms prohibit scraping), ESPN and LPGA websites (no reuse licence), Data Golf (paid licence decision).