About ArmRankings.com
Computed rankings, not opinion lists
ArmRankings.com ranks athletes from recorded match data. The system does not manually place athletes because they are famous, popular, or promoted heavily. Every table is recomputed from division-specific evidence: opponent strength, recency, score, title value, activity, proven wins, elite-schedule context, adjacent-class support, and head-to-head constraints.
The formula is not meant to replace human discussion. It is meant to make the discussion harder to fake. If two people disagree about a ranking, they can point to the same inputs and ask which assumption should change.
Ranking score
Higher score ranks higher. After scoring, recent head-to-head results can reorder close cases when the direct result is strong enough and the score gap is not large enough to justify reversal.
Scoring model
Starting score
Every athlete in a division-arm table starts from a neutral score before match evidence is applied.
Opponent strength
Opponent quality rises when the opponent has wins and falls when the opponent has losses in the same class.
Recency multiplier
Fresh results are amplified. Old results count, but carry less current ranking power.
Match contribution
Wins and losses are weighted by opponent quality and recency. G is score gap. T is title-win bonus. D is defense bonus.
Title bonuses
Champions get protection, but only inside their exact division and arm.
Inactivity penalty
Long inactivity pulls current ranking down without deleting the athlete's history.
Adjacent-class support
For 115kg+ Super Heavyweight tables, strong 115kg results can add limited support when the super-heavyweight resume is thin.
Moved-up activity support
An undefeated title-level winner who moves up can keep lower-class credibility when later same-arm wins in higher classes prove current form. H is proven higher-class wins after the lower-class result.
Evidence constraints
These constraints prevent the common ranking mistakes: overrating one old result, rewarding activity without elite wins, burying an athlete only because he fought elite opponents, or letting a thin resume jump proven contenders.
Thin record penalty
A one-match or two-match resume cannot rank too high unless the evidence is exceptional.
Stale thin resume
Two old results cannot hold a current top ranking without newer elite evidence.
Old one-match resume
A single old result cannot carry a current elite ranking by itself.
Old one-match title result
Former title evidence still counts, but current rankings need current activity.
Unproven win penalty
One win over an unproven opponent should not create a top ranking.
Top-tier gate
Activity alone is not enough. A strong record needs at least one proven same-division win near the podium.
Proven-win bonus
Wins over athletes with proven same-division records are rewarded.
Elite-schedule correction
An athlete with no wins still needs wins to rise, but repeated losses to champions or proven contenders are not treated like ordinary losses.
Competitive elite loss
Taking rounds from a title-level opponent can support the ranking slightly. A shutout loss does not receive this bonus.
Accumulated-loss cap
Many losses still cap the upside, but the extra penalty is reduced when those losses came against elite opposition.
Head-to-head override
If Athlete B is scored above Athlete A, but Athlete A recently beat Athlete B in the same division and arm, the system checks whether the score gap is small enough to respect the direct result. The direct winner stays above when:
| Age of direct result | Condition | Max score gap overridden |
|---|---|---|
| ≤365 days | close direct win | 220 |
| ≤365 days | decisive direct win | 260 |
| ≤730 days | close direct win | 120 |
| ≤730 days | decisive direct win | 160 |
| >730 days | decisive direct win only | 35 |
| ≤1095 days | thin-resume direct win | 220 |
This is why a direct win matters, but does not freeze the ranking forever. Newer elite wins, title results, and stronger activity can eventually overcome an older head-to-head loss.