# Someone Just Created An Algorithm To Predict Survival In The Squid Game’s Glass Bridge

The series is centered around 456 debt-ridden players from different walks of life.

South Korean survival drama television series ‘Squid Game‘ has taken the world by storm ever since it was released in September. The series is centred around 456 debt-ridden players from different walks of life. They are invited to a tournament to play a set of children’s games with deadly outcomes for losing a round in their quest to win a ₩45.6 billion prize.

In episode 7, the 16 players that are still alive play the Glass Bridge game. The rules are simple: 18 steps to the end of the bridge, each step with two glasses the player can step on. One of the glasses is safe, but the other will break, and the player would be ‘eliminated’.

#### THE BELAMY

Alejandro Martinez Vargas, Principal Data Scientist and Decision Science at Atlassian, has put forward theories on the odds of surviving and how many were expected to cross the bridge safely.

Alejandro says each player is expected to ‘unlock’ for the rest of the players. Even if a player steps on the wrong glass, it will unlock 1 step for the next player in line. So that’s a 100% probability of unlocking at least 1 step. Of course, there is a 50% chance to select the safe glass, in which case, the player would also be unlocking the next step (as they will do a second jump). Then, 25% chance (50% x 50%) to have 2 safe jumps (unlocking 3 steps for the rest of the players), a 12.5% chance to have 3 safe jumps, etc. So, we can calculate the expected steps each player will unlock, and it looks like:

Each player is expected to unlock exactly 2 steps. There are 18 steps, so we expect 9 players to unlock them, and therefore we expect 7 players to survive. In the show, only 3 players survived, so they didn’t do very well.

Player 1 would need to pick the right glass 18 times. Each of the jumps has a 50% chance, so the probability of Player 1 reaching the other side is:

That’s 1 in over 260.000! Sadly, not great odds for Player 1.

For the others, their chance of survival depends on how well or how bad the players before them do. The probability of player K reaching step X depends on two things:

• The probability of player K-1 (prior player) reaching step Y (with Y < X )
• The probability of player K taking X-Y steps

As we know that Player 1 had it difficult, we might have thought that Player 5 for sure was going to have much better chances, but not really! Player 10 is the first player with more chances to survive than to be ‘eliminated’.

And this chart also proves what we mentioned earlier: we expect 7 players to survive!

In the past few weeks, the show has earned rave reviews and has become a pop-culture phenomenon with memes, facts, references, and theories on every major social media platform.

## More Great AIM Stories

### Top Countries Pumping Money Into Quantum Computing Technology

Victor is an aspiring Data Scientist & is a Master of Science in Data Science & Big Data Analytics. He is a Researcher, a Data Science Influencer and also an Ex-University Football Player. A keen learner of new developments in Data Science and Artificial Intelligence, he is committed to growing the Data Science community.

## Our Upcoming Events

Conference, in-person (Bangalore)
MachineCon 2022
24th Jun

Conference, Virtual
Deep Learning DevCon 2022
30th Jul

Conference, in-person (Bangalore)
Cypher 2022
21-23rd Sep

### Discord Server

Stay Connected with a larger ecosystem of data science and ML Professionals

### Telegram Channel

Discover special offers, top stories, upcoming events, and more.

##### Most used enterprise tools for digital-first firms in 2022

The enterprise software market got a major fillip after remote working became the norm.

##### AIM turns 10!

We have an opportunity to build a legacy.

##### Artificial bee colony and its applications to optimization problems

Honey bees’ foraging behaviour inspired the development of artificial bee colony.

##### 7 tricks to speed up the training of a neural network

There are a few approaches that can be used to reduce the training time time of neural networks.

##### Why the high accuracy in classification is not always correct?

The high accuracy of classification model could be misleading.

LTI and Mindtree both play in Analytics services businesses, just like most other large IT/ITes service providers. But, what would the analytics services business of the merged entity look like?

##### GitHub now offers math support in markdown

GitHub’s math rendering capability uses MathJax; an open-source, JavaScript-based display engine.

Meta recently organised messaging event called ‘Conversations.’

##### Wipro announces 40,000 sq.ft. Innovation Studio in Texas

The studio will leverage Wipro’s deep reservoir of IPs, patents, and innovation DNA.