Methodology
The GOAT Score is a weighted, normalised composite. The complete model is published so every result can be examined.
Methodology Paper v1.0
The full specification — every weight, multiplier and limitation
Data sources, normalisation, why each sport has the competition-depth multiplier it has, era adjustment, uncertainty, edge cases, known biases and how to file a correction.
Read the paper →Category weighting
GOAT = 0.25·Dominance + 0.20·Peak + 0.20·Longevity + 0.20·Honours + 0.15·Statistics
Worked example · Lionel Messi · in-sport 97.0 · cross-sport 97.0
1 · Collect raw records
Titles, appearances, records and per-season statistics are collected from public, citable sources and stored with a link back to the origin.
- Career goals: 879
- Career assists: 383
- Ballon d'Or: 8
- World Cups: 1
- Sources: FIFA, UEFA, Transfermarkt, Opta
2 · Normalise within the sport
Each metric is rescaled against the best performers in the same sport, so a goal, a major and a grand slam are measured on one 0–100 scale.
- football normalization cohort: 4271 records — the full reference universe
- Career goals 879, measured against the 109 records in that cohort that report this metric
- 107 of 108 rank below → percentile 99.1
- Scaled against the cohort best (924): 95.1 / 100
3 · Adjust for era and context
Era strength, competition depth and schedule length are applied so athletes from different decades are compared on equal terms.
- Competition depth (football): ×1.00
- The depth factor is applied only in ALL SPORTS and other cross-sport views.
4 · Roll up into five categories
The adjusted metrics roll up into dominance, peak, longevity, honours and statistics — the five pillars shown on every profile.
- Dominance: 98 / 100
- Peak: 99 / 100
- Longevity: 92 / 100
- Honours: 97 / 100
- Statistics: 99 / 100
5 · Apply the weighting
The five categories are combined with fixed public weights into one GOAT score, recomputed whenever the underlying data changes.
- In-sport: 0.25 × 98 = 24.50 + 0.20 × 99 = 19.80 + 0.20 × 92 = 18.40 + 0.20 × 97 = 19.40 + 0.15 × 99 = 14.85
- = 97.0 GOAT Score
- Cross-sport: 97.0 × 1.00
- = 97.0 canonical ALL SPORTS score
Per-sport normalisation
| Boxing | Anchor metric · World titles | World titles, era dominance and the quality of the fighters beaten. |
| Swimming | Anchor metric · Olympic golds | Olympic medal hauls, world records and multi-Games dominance in the pool. |
| Football | Anchor metric · Career goals | Club and international output normalised across eras, leagues and defensive contexts. |
| Tennis | Anchor metric · Grand Slam titles | Slam depth, weeks at No. 1 and head-to-head strength inside the strongest fields. |
| Golf | Anchor metric · Major championships | Majors, weeks at world No. 1 and strokes gained relative to the field. |
| Formula 1 | Anchor metric · World titles | Championships, win rate and qualifying pace adjusted for machinery advantage. |
| Esports | Anchor metric · Major trophies | International trophies, rating dominance and longevity in fast-rotating metas. |
| Basketball | Anchor metric · NBA titles | Rings, MVPs and box-score dominance normalised across pace and era. |
| Baseball | Anchor metric · Home runs | Home runs, OPS and two-way value weighed against league context. |
| American Football | Anchor metric · Super Bowl rings | Super Bowl rings, MVP share and volume production across long careers. |
| Ice Hockey | Anchor metric · Career points | Stanley Cups, scoring titles and all-time point totals adjusted for era scoring. |
| Cricket | Anchor metric · Test average | Test average, century counts and World Cup impact across formats. |
| Table Tennis | Anchor metric · World & Olympic titles | Olympic and World titles plus months at world No. 1 in the deepest fields. |
| Volleyball | Anchor metric · Olympic golds | Olympic gold, World Championships and MVP awards indoor and on sand. |
| Athletics | Anchor metric · World records | World records, Olympic and World Championship gold, and margin of dominance over the field. |
Global participation & competition depth
Raw records are normalised within each sport, adjusted for era and context, grouped into five categories, then combined with fixed weights. Global rankings also apply competition depth; legacy mode separately adds measured media and commercial impact.
Cross-sport GOAT = In-sport GOAT Score × Competition Depth
The same canonical calculation powers ALL SPORTS, athlete profiles and this worked example. Inside a sport, the raw In-sport GOAT Score remains unchanged.
Within a sport, the pure 0–100 score is used. Competition depth applies only to global and cross-sport comparisons, with both values shown.
Competition Depth Index — every sub-factor →
| Football | ×1.00 | |
| Basketball | ×0.98 | |
| Swimming | ×0.97 | |
| Boxing | ×0.96 | |
| Athletics | ×0.96 | |
| Cricket | ×0.94 | |
| Tennis | ×0.93 | |
| American Football | ×0.88 | |
| Ice Hockey | ×0.88 | |
| Baseball | ×0.86 | |
| Volleyball | ×0.84 | |
| Golf | ×0.82 | |
| Formula 1 | ×0.80 | |
| Table Tennis | ×0.78 | |
| Esports | ×0.76 |
Data confidence
The confidence figure shown next to every datapoint is the weighted mean of five sub-scores, each 0–100. Open the ⓘ next to any number to see its own sub-scores.
| Source quality | 40% | Tier of the strongest source behind the record: governing body or league (100), established statistical reference (90), other public source (75). |
| Cross-source agreement | 25% | How many independent sources carry the same record. One source scores 78, four or more score 100. |
| Record completeness | 20% | How completely the era is documented. Pre-1930 archives are reconstructions and score lowest; the modern era scores 100. |
| Freshness | 10% | How recently we re-checked the record against its sources. |
| Manual verification | 5% | Whether a human checked the record. Agent-discovered records that nobody has audited score lowest. |
Agent-discovered records that no human has audited are capped at 62% regardless of how strong their sources look.
Ranking eligibility
Incomplete data never produces a public ranking. A record must clear every threshold below before it receives a GOAT Score; everything else is published as "Not yet ranked — pending verification" and is excluded from leaderboards, podiums and sport hubs. Saying we do not have enough reliable data yet is more honest than publishing a number that understates a great career.
- · Record completeness ≥ 80%
- · Data confidence ≥ 75%
- · At least 3 recorded career metrics
- · A season-by-season career series is present
- · The record is editorially verified, not an unaudited public-source entry
| Season-by-season career record | 30% | How much of the career is covered by a year-by-year performance series. Six or more seasons score 100. |
| Core career metrics | 25% | How many core career metrics are recorded. Five or more score 100. |
| Documented honours | 20% | Whether titles, majors, medals or awards are recorded against the athlete. |
| Independent sources | 15% | How many independent sources back the record. Three or more score 100. |
| Career timeline | 10% | Whether the first and last competitive years are known. |
Eligibility ledger
The live effect of the thresholds above, sport by sport. Pending records are in the database and visible on their own profile page, but carry no public score.
5087 ranked · 9287 pending · 14374 records total
| Sport | Ranked | Pending | Share ranked |
|---|---|---|---|
| Baseball | 712 | 688 | 51% |
| Football | 674 | 3597 | 16% |
| Ice Hockey | 477 | 1757 | 21% |
| American Football | 373 | 2406 | 13% |
| Cricket | 372 | 111 | 77% |
| Basketball | 369 | 435 | 46% |
| Athletics | 292 | 16 | 95% |
| Swimming | 267 | 21 | 93% |
| Golf | 259 | 6 | 98% |
| Volleyball | 252 | 51 | 83% |
| Tennis | 248 | 2 | 99% |
| Esports | 241 | 85 | 74% |
| Boxing | 231 | 0 | 100% |
| Table Tennis | 177 | 112 | 61% |
| Formula 1 | 143 | 0 | 100% |
Normalization cohorts
Three distinct cohorts govern the model. Percentiles are always computed against the full normalization cohort of the sport — never against a hand-picked shortlist — while only the eligible cohort receives a published GOAT Score.
- Normalization cohort
- Every record tracked in the sport. All percentile and 0–100 metric rescaling is computed against this full reference universe, not against a hand-picked shortlist.
- Verified cohort
- Records that have passed editorial review with full source attribution. Agent-discovered records stay outside this cohort until a human verifies the underlying facts.
- Eligible cohort
- Verified records that also clear the completeness and confidence thresholds. Only this cohort receives a published GOAT Score and a place in any ranking.
14374 normalization · 5206 verified · 5087 eligible
| Sport | Normalization | Verified | Eligible | Data coverage |
|---|---|---|---|---|
| Football | 4271 | 696 | 674 | 16% |
| American Football | 2779 | 381 | 373 | 13% |
| Ice Hockey | 2234 | 480 | 477 | 21% |
| Baseball | 1400 | 714 | 712 | 51% |
| Basketball | 804 | 387 | 369 | 46% |
| Cricket | 483 | 404 | 372 | 77% |
| Esports | 326 | 245 | 241 | 74% |
| Athletics | 308 | 294 | 292 | 95% |
| Volleyball | 303 | 259 | 252 | 83% |
| Table Tennis | 289 | 185 | 177 | 61% |
| Swimming | 288 | 273 | 267 | 93% |
| Golf | 265 | 265 | 259 | 98% |
| Tennis | 250 | 249 | 248 | 99% |
| Boxing | 231 | 231 | 231 | 100% |
| Formula 1 | 143 | 143 | 143 | 100% |
Cultural & brand impact
The 1–100 cultural index combines social reach, press presence, commercial value, brand strength and encyclopedic footprint.
Total legacy = Cross-sport GOAT × 82% + Cultural & Brand Impact × 18%
Pure Sporting Data is the default. Total Legacy & Brand separately blends in measured cultural impact, without changing sporting credit.
Career Day, provenance & versioning
Career Day aligns athletes by the exact number of days since professional debut, so different eras are compared on the same career clock.
Currency, participation and deduplication gates prevent fabricated ratios and double counting. Only verified sources reach public rankings; immutable snapshots preserve history.
calc calc-v1 · eligibility elig-v1 · methodology method-v1
Data transparency
Every athlete profile lists its public sources. Records are point-in-time snapshots, and active careers continue to change.