Game of Thrones final season deaths as predicted by algorithm

Who will die in the final season of Game of Thrones?

A data scientist working for a Boston-based outfit called DataRobot wrote a complex algorithm to try and predict which favorite characters are most likely to perish in season 8 (results below).

Taylor Larkin studied 2,000 living and dead A Song of Ice and Fire characters detailed on a Wiki fanbase and analyzed their traits — such as their gender (female characters, it turns out, die less frequently than males, who are more likely to be in battles), age (older characters are more vulnerable), house (Targaryens, in particular, haven’t fared well), nobility status, and how many of each character’s relatives have died (which turned out to be a very pivotal trait given how many families on the show have been decimated by rivals).

Using automated machine learning, Larkin tried to use the past events in Westeros to project which of the currently surviving characters will not survive the show’s final six episodes.

Below are the results.

Whether the following requires a spoiler alert is, of course, unknown. I guess it depends on how much you trust AI. Perhaps that’s a better question for Westworld fans? A computer trying to reveal GoT spoilers … this must be what Elon Musk was warning us about!

• Daenerys Targaryen – 83.77% chance of death

• Jaime Lannister – 72.91% chance of death

• Tyrion Lannister – 70.76% chance of death

• Bran Stark – 66.02% chance of death

• Cersei Lannister – 60.39% chance of death

• Jon Snow – 58.99% chance of death

• Euron Greyjoy – 54.95% chance of death

• Sansa Stark – 50.28% chance of death

• Arya Stark – 49.04% chance of death

• Gendry – 39.87% chance of death

Of course, if there’s one thing that GoT has consistently shown is that its twists are unpredictable. We’ll all find out when Game of Thrones returns in 2019.