Data predicts who will die in Season 6, and who should still be alive

Those who would love a deep dive into A Song of Ice and Fire prior to Game of Thrones Season 6 have a huge problem: there’s nothing to dive into. With no books, no guide, no map, and no ability to know what will happen this year, we’re starting to see the mathematically inclined among us try and desperately predict the unpredictable. We’ve already had one study that proved Tyrion is the main hero of the series (and therefore, logically, should survive.) Now, “A Song of Ice and Data” has come along to predict the survival chances of the remaining characters this season, as well as prove once and for all that Jon Snow’s death is a fake out.

Taking data from all credible online sources covering the source material, including the indispensable A Wiki of Ice and Fire and the Game of Thrones wiki, the program’s algorithm categorizes every character on the show and in the books—2,028 in all—by such factors as gender, age, title, popularity (by pageviews and Google searches) as well as how many of their family members have already been offed.

Some fun stats:

• The male:female ratio in the novels is over 2:1
• But when it comes to survival, women survive by just under 2:1 ratio—by the end of the last book, only 43% of the male characters have survived, compared to 79% of women.
• Men are usually nobles, while women are more likely to be peasants.
• Book 2 (A Clash of Kings) has the most characters: 437
• Book 1 (A Game of Thrones) has the most male characters: 271 (to 115 women)
• Book 5 (A Dance With Dragons) has the least amount of characters appearing in it, by a huge margin: 104. That’s in comparison to the other four novels, all of which have averaged over 350 characters per installment. (AGoT: 386, ACoK: 437, ASoS: 356, AFfC: 391.)
• There are typically about 30 characters per episode.

Ok, now for what you really came for—who is most likely to die in Season 6/The Winds of Winter, according to this study?

• Tommen Baratheon (97%)
• Stannis Baratheon (96%)
• Daenerys Targaryen (95%)
• Davos Seaworth (91%)
• Petyr Baelish (91%)

Tommen is no surprise. Gold will be their crowns and gold will be their shrouds. And Cersei has been losing an average of one child per season. Tommen is so next. Stannis is surprising, until you remember this also includes the books—the show just got there first. We’ve been predicting that Davos wouldn’t make it to the end of the season, based on our own algorithm of how many interviews Liam Cunningham has been doing during the promotional period. I like Davos, though—I hope we’re all wrong.

Littlefinger’s ranking being so high is startling, but not as startling as Dany’s. Now, I know we’ve joked here that she’s going to die in some freak accident, just before crossing the Narrow Sea, and totally disappointing everyone. But we should also point out that if Dany were a regular human, she would have died at the end of Season 1, when she stepped inside a roaring bonfire. Surviving death is just one of those things she does. Littlefinger is also an ‘against the odds’ survivor.

But it’s the “Who is most likely to survive” calculations where things get really interesting:

• Sansa Stark (3%)
• Jon Snow (11%)
• Cersei Lannister (16%)
• Mace Tyrell (18%)
• Roose Bolton (28%)

Sorry, Sansa haters. She’s the most likely to survive out of everyone, because so many of her family members are gone, and because of her own luck (if one can call it that) in getting away from psycho husbands. Same with Cersei—as one of the very few Lannisters left in King’s Landing, she looks to be fated to live a long time, alone with her failures. I like that Mace will survive because he’s a bumbling fool. I’m a little surprised about Roose, but if he can outlive Ramsay, I think that might hold.

But as you all can see, the one sitting at Number 2 on the list gives the game away. Jon Snow’s stats say he should not have died. According to these calculations, the character’s popularity keeps him safe, despite the fact that he…y’know, already died. Frankly, the researchers just don’t believe it. Numbers don’t lie.