The BSB MECCA Index Methodology

How we calculate every athlete grade — every pillar, weight, data source, and the edge cases. We publish this in full because a power ranking is only as credible as the maths behind it.

The short version

Each athlete is scored across multiple pillars of public-domain influence. Each pillar is z-scored within the active rider population, weighted, and combined into a composite. The composite is mapped to a percentile-based letter grade from A+ (top 1%) to F (bottom 5%).

No subjective opinion. No team allegiances. No paid placement. Same maths applied identically to every athlete.

Pillars

The v1 grade combines two pillars currently. Three more are launching with BSB MECCA Pro and the broader rollout. Weights renormalise across active pillars per athlete — missing data on one pillar doesn't penalise an athlete unfairly.

PillarWeightStatusSource
Search demand35%LiveAhrefs (Google + Bing global monthly search volume)
Wikipedia interest15%LiveWikipedia REST API (article pageviews, last 30 days)
Social reach30%LiveCombined verified follower counts across Instagram, TikTok, X and YouTube
Press tier-weighted mentions10%Phase 2NewsAPI / GNews, classified by source domain rating (Tier 1 = DR 80+, weighted 5×)
LLM Share-of-Voice10%Phase 2Mentions across ChatGPT / Gemini / Perplexity / Claude when answering superbike questions
Search demand35%

Global monthly search volume for an athlete's name across Google and Bing. Pulled from Ahrefs Keywords Explorer. Updated weekly.

Why log-scaled: raw search volume is heavy-tailed (Jett Lawrence has 60k searches; an EMX125 amateur may have 90). Without log transform, the top 3 names crowd out everyone else. We use log10 so the gap between #1 and #10 is meaningful, not exponential.

Namesake correction: some athletes share a name with a politician, actor, or athlete from another sport (e.g. James Stewart the politician/baseballer/footballer is the same name as the superbike legend). For these riders we substitute the disambiguated MX-only volume — sum of "{name} superbike", "{name} superbike", and known nicknames (e.g. "Bubba Stewart"). This is documented per-rider in the Pro view.

Wikipedia interest15%

Monthly pageviews of an athlete's English Wikipedia article. Pulled from Wikimedia REST API. A near-perfect proxy for sustained encyclopedic interest — riders who are merely active on socials don't accumulate Wikipedia views. Riders who are "the GOAT" do.

Riders without a Wikipedia article get a zero on this pillar — which is expected for amateurs and emerging pros. Their grade is calculated from the other pillars they do have data on.

Social reach30%

Combined follower count across Instagram, TikTok, X (Twitter) and YouTube. Pulled weekly from public profile pages and stored as total reach. Higher reach scores higher; the log-transform stops a few mega-accounts crowding everyone else out.

Phase 2 adds engagement rate (a 100k follower account with 0.5% engagement scoring lower than a 50k account with 6%) and a recency factor (zombie accounts lose 30% of their social score).

Press tier-weighted mentions10% (Phase 2)

Mentions of an athlete in news media, classified by source domain authority. Tier 1 (DR 80+ — BBC, ESPN, Reuters) weighted 5×. Tier 2 (DR 50-79 — Speedweek, Cycle News, MX Vice) weighted 2×. Tier 3 (regional / niche / blog) weighted 1×.

Last 90 days + lifetime totals stored separately. Pro subscribers see the breakdown.

LLM Share-of-Voice10% (Phase 2)

When ChatGPT, Claude, Gemini, or Perplexity answer prompts like "best superbike riders right now", "top superbike riders 2026", or "greatest superbike of all time" — who gets named? The percentage of cited mentions per athlete across these LLM outputs.

Why this matters: an athlete's presence in AI training data and search-grounded LLM answers is increasingly how the next generation of fans first encounters them. It's the future of brand awareness.

How the composite score is calculated

For each athlete with data on at least one pillar:

  1. Each pillar's raw value is log-transformed (log10) to handle heavy-tailed distributions
  2. The log values are z-scored across the entire active rider population for that pillar
  3. A composite z-score is produced by weighted sum of the athlete's pillar z-scores
  4. Weights are renormalised to active pillars only — missing data does not penalise the athlete
  5. All composite scores are ranked, and percentile bands map to letter grades
composite_z = Σ (pillar_z × pillar_weight) / Σ pillar_weight (active pillars only)

Grade bands (population percentile)

A+Top 1%
ATop 5%
A−Top 10%
B+Top 25%
BTop 50%
B−Top 75%
CTop 90%
DTop 95%
FBottom 5%

Refresh cadence

  • Search demand: weekly
  • Wikipedia interest: weekly (rolling 30-day window)
  • Social reach: weekly (Phase 1b onward)
  • Press tier mentions: daily (Phase 2)
  • LLM Share-of-Voice: bi-weekly (Phase 2)

Grades are recalculated nightly. Athletes can move grade bands as their data shifts.

Eligibility

To appear on the public leaderboard, an athlete must have meaningful data on at least one pillar (currently: a search-volume value greater than zero). Athletes without sufficient public signal are not graded — better to be silent than guess.

Verified athletes (those who have claimed their BSB MECCA profile) display a verified badge and may have richer data sourced from authorised social accounts.

Edge cases we explicitly correct for

  • Namesakes — riders sharing a name with a politician, actor, or athlete from another sport get disambiguated search volumes. List of corrections is reviewed quarterly.
  • Retired athletes — historical legends (Carmichael, McGrath, Stewart) compete against current athletes. We don't apply a retirement decay because cultural relevance is the signal we're measuring, not active competition. Active-only filtered views are available with BSB MECCA Pro.
  • Duplicate rider records — our database aggregates Euro and US scraping sources; some riders have duplicate slugs from name variants. Only riders with curated search-pillar data appear on the leaderboard, which removes the noise.
  • Bot-inflated socials — once social pillar is live, accounts with abnormally low engagement-to-follower ratios are flagged and capped to prevent gaming.

What this is not

  • It is not a "best rider" ranking. BSB MECCA Power Rankings measure on-track performance via ELO. The BSB MECCA Index measures combined on-track and off-track influence — a different question.
  • It is not editorial. No journalist input, no team affiliation. Only public-domain data and transparent maths.
  • It is not static. Grades update weekly. Hot riders climb. Quiet ones drop.

Got a question?

Methodology questions, corrections, or partnership enquiries: email [email protected].

See your favourite rider's full breakdown.

BSB MECCA Pro launches with per-pillar sub-scores, 90-day trend graphs, peer comparison, and rising-star alerts. Join the waitlist →


Explore the database

Results verified against multiple sources; gaps exist in older records. Spotted something wrong? Let us know, we update fast.