Squid Game - Survival Analysis (Kaplan-Meier)

The Korean show Squid Game was a great success in 2021. The show follows to 456 persons who decided to be part of a game in which after six different stages, in which the players participate in different children games, only one can the winner who will get lots of money. 15 players quit and the other 440 died along the way (Because if you do not win, you die. You may know that by now). Inspired by the game I decided to implement a Kaplan-Meier curve to determine the probability of survival at different stages of the game.

The Kaplan-Meier is a survival analysis non-parametric method. It allows determining the probability of survival of individuals of a group after a certain time from the occurrence of an event. This method is commonly used in medical research to determine survival potential after organ transplantation. But it can be used in a lot of other different fields.

Here I have implemented the Kaplan-Meier curve for the participants in the Squid Game series. The events that have taken in account are listed below:

Event	Event number
First game: Red light, Green light	1
Participants decided not to continue, but come back	2
Second game: Honeycomb	3
Riot at night	4
Third game: Tug of War	5
Doctor execution	6
Fourth game: Marbles	7
Suicide	8
Fifth game: Hopscotch	9
Dinner before last game	10
Sixth game: Squid Game	11

The participants that 'quit' the game including player number 1, have been labeled as censored. Here are the results.

Conclusions

The probability of survival, of each participant, from the start of the game was very low 1 in 456 , that is nearly 0.23%.

However, if you survived the first game and decided to come back your chances are doubled to 1 in 187, which is nearly 0.53 %. Still pretty low if you ask me.

Looking at this graph, what would you do? Would you play?