top of page

Game Scores

Writer's picture: Clint MortClint Mort

Updated: Jun 15, 2024



A player's game score can be calculated by taking their aggregate percentile rankings in the six key performance measurements and weighting them, to create a composite score that reflects their overall performance in a particular season of Survivor. The six performance measures are as follows:



Correlation Weights:

To calculate the weighting for each of the performance measurements, every player's game ranking for each category is compared to the final placement ranking for that player in their respective seasons. The correlation co-efficient is calculated, which measures the strength of correlation between the final placement rankings and the six performance measurements.


The correlation co-efficient for all six performance measurements can be added together to calculate their individual proportion of the total. The percentage is then calculated to get applied weights. The higher the correlation is between a performance measure and the final placement rankings, the more impact that measurement will have on the player's final game score.


For example, the vote avoidance measurement has a correlation co-efficient of .7320 to final placement rank. This co-efficient makes up 22.55% of the total (.7320 / 3.2464). Therefore, a player's percentile rank for vote avoidance will constitute 22.55% of their overall game score.


Game Score Calculation:

To calculate a player's game score, a player's percentile rank is taken for each of the performance measurements and converted to a whole number (% x 100). Then each percentile rank is multiplied by the weighted portion of the correlation co-efficient. By adding the six weighted whole numbers together, a final weighted score is determined. For example, Tyson Apostol (winner of season 27) had a weighted game score of 89.9.


  • Vote Cluster percentile rank = 79.2% x 100 = 79.2 x 11.64% = 9.2189

  • Vote Accuracy percentile rank = 64.6% x 100 = 64.6 x 12.99% = 8.3915

  • Challenge Performance percentile rank = 89.4% x 100 = 89.4 x 0.61% = .5453

  • Immunity Idols percentile rank = 99.6% x 100 = 99.6 x 25.66% = 25.5574

  • Vote Avoidance percentile rank = 90.6% x 100 = 90.6 x 22.55% = 20.4303

  • Safety Earned percentile rank = 97.1% x 100 = 97.1 x 26.54% = 25.7703


Weighted game score = 9.2189 + 8.3915 + .5453 + 25.5574 + 20.4303 + 25.7703 = 89.9137

In the first ten seasons, there were no immunity idols available in the game. In addition, occasionally players are eliminated from a season before any idols are found or activated. For example, Bruce Perreault was removed from season 44 approximately twelve hours into the game due to a head injury suffered at the first challenge. Even though there would eventually be nine idols in season 44, Bruce was eliminated before any were active. Therefore, in cases where players didn't play in the game where there was an active idol, an idol score is not considered.


For example, Sandra Diaz-Twine (winner of season 7) had a weighted game score of 93.1.


  • Vote Cluster percentile rank = 87.4% x 100 = 87.4 x 15.66% = 13.6868

  • Vote Accuracy percentile rank = 87.2% x 100 = 87.2 x 17.48% = 15.2426

  • Challenge Performance percentile rank = 2.5% x 100 = 2.5 x 0.82% = .0205

  • Vote Avoidance percentile rank = 99.6% x 100 = 99.6 x 30.33% = 30.2087

  • Safety Earned percentile rank = 95.1% x 100 = 95.1 x 35.71% = 33.9602


Weighted game score = 13.6868 + 15.2426 + .0205 + 30.2087 + 33.9602 = 93.1188


All individual game scores were calculated and ranked against the entire population of game scores. A percentile rank was calculated for the ordinal rank. Sandra Diaz-Twine's season 7 game score of 93.1 ranks her 10th overall out of 839 total game scores, which fell into the 98.9 percentile. Tyson Apostol's season 27 game score of 89.9 ranks him 19th overall out of 839 total game scores, which fell into the 97.8 percentile.


A final correlation co-efficient was calculated for the game score rank against the final placement ranking, which equaled .8835. This shows that by combining the measurements by weight, the correlation between final placement and game score rank is increased closer to 1, and at a level to sufficiently be considered highly correlated.


To view a full list of game scores, please click button below:


*** updated through season 46 ***

40 views1 comment

Recent Posts

See All

1 Comment


©2022 by Survivor Analytics. Proudly created with Wix.com

bottom of page