Spoiler alert: This story provides no plot details from past or future episodes of “Game of Thrones,” but if you want to stay totally in the dark about the speculation, read no further than the fourth paragraph.
“Game of Thrones,” the HBO sword-sorcery-and-sex series that begins its seventh season tonight, has spawned a geekier kind of game: doing data analytics to study who’s on top and who’ll get the ax (figuratively and literally).
The Wall Street Journal delved deeply into the data geekery last week, noting that fans have recorded the screen time and other statistics for nearly 200 notable characters over the course of the past 60 episodes. Intricate algorithms seek to correlate the characters’ characteristics with their chances of surviving the poison pens wielded by book author George R.R. Martin and the series’ screenwriters.
Indian data scientist Shail Deliwala ran the numbers and came up with several correlations. Some mesh with common sense (for example, characters with lots of dead relatives are less likely to survive), while others mesh more with literary sense (for example, the larger the attacking force, the higher the chances that the defenders will prevail).
Another analysis, conducted by Gowri Shankar, identified the most pugnacious king and the winningest fighter on “Game of Thrones.”
So what do the data say about the seventh season? Looker Data Sciences’ Kelly Payne lays out some not-totally-scientific guesses based on past trends. These guesses are not based on inside information, so technically, you couldn’t call them spoilers. Nevertheless, read on at your own risk:
Admittedly, these guesses are pretty much in line with what “Game of Thrones” fans would have predicted, based on how things were shaping up in Season 6. Nevertheless, the episodes ahead should provide some opportunities to put data science to a trial by fire … and ice.