Using Math to play the Game of Thrones: Which characters will be killed off Next ?

the game of throne math

the game of throne math

HBO’s Game of Thrones is one of the most captivating series of this generation and several fans are on the look out to guess what’s coming next from the super-violent, and extremely-sexual fantasy filled novels written by George R.R. Martin. Of course most of it is just sheer guess-work on a close reading of the books—and definitely some vivid imagination—but a one Richard Vale a teacher at the University of Canterbury wants to use some mathematical logic to give us the answers we seek.

Richard uses Baysian statistics (don’t worry if you are lost). This is a field of mathematics used to evaluate the probability of future events. This is a much needed prediction given this series keeps surprising everyone, especially with the deaths of popular characters. The Richard Vale has released a first ever mathematics paper that will endeavor to predict which characters wont survive next books of the novel, so brace yourself because the paper includes a lot of spoilers in its abstract. The statistics teacher seeks to predict which characters will receive the most attention in the author’s next two Game of Thrones books and he critically examines the number of chapters devoted to each of them in the previous five books.

FROM OUR SPONSOR- Continue for more content

In what began as an explanatory tool for his students, Vale wondered whether he could find out anything interesting by modeling it while also having fun with his students.

Jon Snow may not survive

It’s all based on a simple idea. Every character in Game of Thrones is told from the point of view of another specific character, and Vale used the number of chapters dedicated to each character in each book to create a simple mathematical model to prognosticate how many chapters might be dedicated to each character in the upcoming books. Using this method obviously can’t predict specific storyline and plot twists. But it endeavors to allow for some guesstimates.

Vale says

“Presumably, dead implies zero POV chapters,So there should be a small amount of information about the potential deaths of characters if we believe the model. I am cautiously pessimistic about the model’s chances of giving a good prediction,”

Lets look at an example, in his predictions, Jon Snow’s character has a probability of 0.38 hence having zero chapters in the sixth book, and the probability of him having zero chapters in the seventh book a little over 0.67. In a nut shell Jon Snow won’t survive the end of the sixth book, if we base ourselves entirely on his model.

The death of Joffrey Baratheon
The death of Joffrey Baratheon

Data Paucity

With over 5,216 pages and five books for Vale to use, it’s surprising that there’s not much information available. The prediction ‘algorithm’ gives no obvious predictable pattern to the several chapters any other individual character will drive before being killed off. The other down side is that the model doesn’t factor in take the content of the previous books. This alone may be the biggest issue with the math being used.

Vale goes on to say in his paper that:

“In general, the best predictions are obtained by a combination of modelling and common sense, “Here we focus entirely on the modeling side and leave common sense behind.”

This means it’s possible that there will be chapters dedicated to characters already dead, for instance, and it says some characters, who are clearly alive, may not appear in any chapters.

The beauty of Bayesian statistics is that it gives room for you to update predictions if and when it’s becomes available.  But he doubts the model will actually do well enough for him to bother updating it with fresh data. With the rumors that Martin will desert the practice of writing each chapter from a different character’s point of view, could break this model he has used.

What do we learn from this paper? While the paper might not help you win any awards or betting, it does help get a sense of how mathematicians draw close to predictions. And now we can wait and see if this model pays off. You can read the paper in the source link below.


Source: Richard Vale Mathematics paper