Three days into the CS2Predict Rio Major prediction contest, 18,400 participants have registered — 43% above the projected participation baseline. Brazilian community dominates the leaderboard (52% of top-100 predictors are Brazilian). Prior Cologne winner "donk_lover_UB" (Mongolian) currently defends the crown at #1 overall.
Participation metrics
- Total registered participants: 18,400 (target 18-25K)
- Country breakdown: Brazil 34%, Mongolia 18%, US 14%, Europe 22%, other 12%
- Predictions submitted (first 3 days): 47,600 across 24 EU RMR matches + 20 Americas RMR matches + 14 Asia RMR matches (58 total match cards active).
- Registration method: 68% email-only, 32% Steam authentication.
Top-10 leaderboard (as of July 24 evening)
- donk_lover_UB (Mongolia, prior Cologne contest winner) — 89 pts
- fallen_apostle (Brazil, Rio-based analyst) — 87 pts
- arT_disciple (Brazil) — 85 pts
- zywoo_mvp (France, active during Vitality bootcamp cycle) — 84 pts
- navi_kingfisher_stan (Ukraine, joined post-kingfisher signing) — 82 pts
- silverside (US, tier-2 analyst) — 81 pts
- chopper_iq (Russia, Spirit-adjacent) — 80 pts
- biguzera_ftw (Brazil, paiN-adjacent) — 79 pts
- furia_rio_gold (Brazil, FURIA-adjacent) — 78 pts
- kscerato_style (Brazil) — 77 pts
The Brazilian community dominance
Brazilian participants make up 52% of top-100 despite representing 34% of total registrations — an over-index of 1.5×. Driver: local knowledge of tier-2 Brazilian teams (MIBR, RED Canids, Legacy, Nouns) that dominate the early Americas RMR match cards. AI model has less training data on tier-2 Brazilian rosters, giving Brazilian human predictors structural advantage in the preseason phase.
The AI baseline performance
- CS2Predict AI baseline score: 71 points (currently at approx position #248 in leaderboard rank).
- AI moneyline accuracy first 3 days: 74% (above YTD 68.4% baseline — post-retraining improvements showing up).
- Human vs AI comparison: Top-10 humans outperforming AI by 8-18 points across the first 3 days. Historic pattern: humans over-perform AI in preseason phases where regional tier-2 knowledge matters.
Historic community winner insight
donk_lover_UB (Cologne 2026 winner) shared his Rio Major approach in an interview:
"Cologne I got lucky on the round-count predictions for Nuke close-out sequences. Rio Major I'm doing something different — I'm cross-referencing bookmaker odds with CS2Predict AI probabilities. When they diverge by more than 8 points on a specific market, I position aggressively on the correct side. If bookmakers and AI converge, I take AI position. Currently up 89 points via that method — good early sign but 60 days to go."
Rewards leaderboard progression
- Top-1 (currently donk_lover_UB): Rio Major sticker capsule 500 units (~$700 retail) + attendance ticket if transferable.
- Top-2-10: Rio Major sticker capsule 100 units each.
- Top-11-25: Rio Major sticker capsule 50 units.
- Top-26-100: Rio Major sticker capsule 10 units.
- Total prize pool value: ~$18,000.
Full leaderboard available at cs2predict.gg.
