# Survivor ELO

> survivorelo.com is the definitive quantitative rating system for Survivor contestants. It applies Elo rating methodology to measure strategic, social, and physical performance across all US and Australian Survivor seasons.

## What This Site Does

Survivor ELO rates every Survivor contestant using two independent Elo-based models:

- **SHALLOW** (Survivor Hierarchical Algorithm for Logistic Likelihood & Odds Weighting) measures strategic and social positioning through vote accuracy, tribal council survival, final tribal council jury votes, and fire-making challenges. K-factor: 8. All players start at 1500.
- **BEAST** (Binary Elo Algorithm for Survivor Trials) measures physical and competitive performance through individual challenge wins, team challenge wins, and Redemption Island duels. K-factor: 16 (individual), 4 (team).

The dataset covers 76 seasons across US and Australian Survivor, rating 968 total contestants (572 established, 396 provisional with fewer than 5 tribal councils attended).

## Key Data

### All-Time Top 10 Players by Strategic ELO (SHALLOW)

1. Parvati Shallow - 1772.6 ELO, S-tier, 5 seasons, 47 tribals, 83.7% vote accuracy
2. Ben Driebergen - 1715.5 ELO, S-tier, 2 seasons
3. Jeremy Collins - 1705.3 ELO, S-tier, 3 seasons
4. Cirie Fields - 1705.2 ELO, S-tier, 6 seasons, 44 tribals, 78.6% vote accuracy
5. Tony Vlachos - 1702.5 ELO, S-tier, 4 seasons
6. John Cochran - 1694.8 ELO, S-tier, 2 seasons
7. Sarah Lacina - 1692.4 ELO, S-tier, 3 seasons
8. Kristie Bennet - 1692.0 ELO, S-tier, 1 season (Australian Survivor)
9. Earl Cole - 1682.1 ELO, S-tier, 1 season
10. Rob Mariano (Boston Rob) - 1681.3 ELO, S-tier, 5 seasons

### Tier System

Players are ranked into tiers based on statistical z-scores (standard deviations from mean):
- S-tier (Sole Survivor): z >= 2.0 - top ~2% of all players
- A-tier (Finalist): z 1.0 to 2.0 - top ~15%
- B-tier (Merge): z 0 to 1.0 - above average
- C-tier (Jury): z -1.0 to 0 - below average
- D-tier (Pre-Merge): z < -1.0 - bottom ~15%

## Site Pages

- [Rankings](https://survivorelo.com/rankings): Complete player rankings with tier filtering, search, and season/version filters. Dual mode: SHALLOW (strategic) and BEAST (challenge).
- [Episode Snapshot](https://survivorelo.com/episode): Per-episode ELO changes at each tribal council.
- [Season Snapshot](https://survivorelo.com/season): Full season player progression and ELO trajectories.
- [Tribe Snapshot](https://survivorelo.com/tribe): Tribe-level performance breakdowns.
- [Compare](https://survivorelo.com/compare): Head-to-head player comparison tool.
- [Blog](https://survivorelo.com/blog): Episode-by-episode ELO recaps and analysis for current seasons.
- [About](https://survivorelo.com/about): Full methodology explanation for both SHALLOW and BEAST models.
- [Player Profiles](https://survivorelo.com/players): Statistical profile pages for every established contestant, at /players/{name-slug}.
- [FAQ](https://survivorelo.com/faq): Direct answers to common questions about the rankings.
- [SHALLOW Methodology](https://survivorelo.com/methodology): How the strategic rating model works.
- [BEAST Methodology](https://survivorelo.com/methodology/beast): How the challenge rating model works.

## Blog Posts

Episode-by-episode analysis is published for current seasons with ELO movement breakdowns:
- Survivor 50 coverage: Episodes 1-5 (ongoing as of March 2026)

## Methodology Summary

Both models use standard Elo rating math: rating changes depend on the outcome and the expectation (upsets cause larger swings). Ratings are zero-sum. Players carry ratings across seasons as returning players. Players with fewer than 5 tribal council attendances are marked provisional and excluded from official rankings.

SHALLOW processes four event types: vote accuracy (correct voters beat incorrect voters), survival (survivors beat eliminated players), final tribal council (jury votes as pairwise matchups), and fire-making (direct head-to-head). The K-factor of 8 was empirically optimized using Brier score across all seasons.

BEAST processes individual challenges (winner vs. all competitors), team challenges (winning tribe members vs. losing tribe members, reduced K-factor), and Redemption Island duels.

## Data Source

Vote, challenge, and boot order data sourced from the survivoR R package (https://github.com/doehm/survivoR).

## Contact

Website: https://survivorelo.com