Iron Score & Iron Signature

The Iron Score

The Iron Score is a cumulative, placement-weighted career rating. Unlike a classic chess ELO — which punishes a champion brutally for a single off-day and would have ranked a 10-time Ms. Olympia below a one-show winner — the Iron Score rewards career depth. Every documented placement adds points; bigger stages and more recent results count for more.

Inputs (all public, all sourced)

• Placement points — 1st = 100, 2nd = 70, 3rd = 50, descending to 3 for deep placements.
• Tier multiplier — apex shows (Ms. Olympia, Rising Phoenix, Arnold Classic) ×3; IFBB Pro ×2; amateur qualifiers ×1.3; regional amateurs ×1.
• Recency weight — 1.0 for the current year, decaying gently to a 0.5 floor for results 20+ years old, so modern athletes and 1980s legends both rank credibly.
• Cultural footprint (Combined board only) — a capped, log-scaled signal from documented social reach, Hall-of-Fame status, authored books and screen appearances. The placement-only board ignores this entirely.

The formula

Iron Score = 1200
  + Σ ( placement_points × tier_multiplier × recency_weight )   ← career placements
  + cultural_footprint  (capped, Combined board only)

Reading the number: the percentile

The raw Iron Score (typically ~1,200 to ~5,500) is only meaningful relative to the field, so every profile leads with a percentile — “Top 3% of ranked athletes”. That is the share of ranked athletes this athlete out-scores. An athlete needs at least 5 head-to-head matchups (i.e. she has shared a stage with 5+ rated athletes) to leave provisional status and appear on the public leaderboard; her profile still shows the score before then, flagged provisional.

What the recency weight means for the all-time question

Because every placement is recency-weighted, the leaderboard is era-weighted by design: a current champion with a deep record will out-rank a legend of the 1990s whose record was just as dominant in her own time. That is deliberate — the Iron Score answers “who matters most in the sport's living record”, not “who was the greatest ever, era-adjusted”. No formula can settle cross-era debates, and we don't claim this one does. For era-scoped views, use the decade filter on the leaderboard; for the raw competition record without cultural signal, use the Placement board.

Three boards, one engine

The leaderboard exposes three views of the same computation: Placement (competition record only), Cultural (reach + legacy, with at least one verified competition record), and Combined (both, with a half-strength boost for whichever signal dominates).

The Iron Signature

The Iron Signature badge names an athlete's standout body part — “Signature: Back”, “Signature: Glutes”. In a sport judged on muscular development, an athlete's calling card is where she dominates, and that is exactly what the badge communicates at a glance.

It is a derived signal, not a tape measurement. Tape measurements are almost never public for female bodybuilders, so we infer dominance from data we can verify: which muscle groups are most decisive in the divisions where she places highest. Concretely, each of her results contributes division_judging_weight(muscle) × placement_strength; summed and normalised 0–100 per muscle, the top score becomes her signature.

dominance(muscle) = normalise(
    Σ over results of  judging_weight(division, muscle) × placement_strength(place)
)
Iron Signature = argmax( dominance(muscle) )   (only shown when score ≥ 50/100)

Because it is inferred from judging emphasis rather than measured, the Iron Signature reflects how an athlete competes and places, not a clinical anthropometric reading. Where richer signals exist — verified public statements, podcast mentions, the rare published measurement — they feed the same per-muscle score (see the broader Methodology).

Known limitations

• Coverage is uneven across eras and federations — an athlete is only as well-rated as the results we have documented for her.
• The Iron Signature follows divisional judging emphasis; two athletes in the same division will trend toward similar signatures unless other signals differentiate them.
• Both metrics recompute nightly. A new result can shift a score the next day.
• Found an error? Submit a correction with a source URL.