What started as a simple Slack post with the new GoT trailer evolved into something far greater, especially for TV fans (and data nerds) here at Samba. Behold our beautiful minds as we debate the movies in which Sean Bean dies and the assumptions to the impact on the narrative.
[11:18 PM] 
Ned Stark rules. The show died with him.


[11:23 PM]
he died?


[8:08 AM] 
well, i’ll have to forget the above comment if I ever decide to binge-watch game of thrones once the series is over.


[9:01 AM]
Me too. But I can at least gather that I should root for team Stark.


[11:01 AM]
It’s not a spoiler if it’s Sean Bean.


[11:01 AM]
because he dies in EVERYTHING?


[11:02 AM]


[11:04 AM]
ok but…


[11:12 AM]
Does he die in Ronin? I can’t remember.
Gah, that movie is so good.


[11:13 AM]
He doesn’t


[11:14 AM]
That’s right, he just takes his shameful payout.


[11:14 AM]
Apparently the ratio is more lives than deaths, 21 to 50.
But Data Science could argue in favor of “the weight of each movie on his career”


[11:14 AM]
or weight to “impact of death on the narrative”.


[11:19 AM]
And you also have to consider a control group. The percentage of deaths to total, we’ll call d2t, is probably 2% or less for the average actor. It’s 42% for him. That’s an anomaly.


[11:21 AM]
I would like to see an estimated survival curve for Sean Bean


[11:23 AM]
Is that actually an anomaly?  I mean, probably, it definitely sounds like one, but someone dies in almost every movie.  What’s the median d2t?


[11:24 AM]
For Zombies that’s probably around 99.9999%


[11:24 AM]
Zombies are already dead; they don’t die *during* the movie unless they’re a main character.


[11:24 AM]


[11:34 AM]
To clarify, is it going to be length of time from the start of the movie that he is expected to survive? Easy peasy.


[11:34 AM]
I like it!


[11:37 AM]
I love how the actor who has died the most is named John Hurt. Fitting.


[11:44 AM]
It doesn’t appear that anybody has compiled his survival times for the movies he died in. For the movies he didn’t die, I can simply used a censored datapoint at the length of the movie though, so I only need to watch the ones where he did die (I think 21 movies?) to find out where he did die. I can also probably drastically reduce my watch time by fast-forwarding to likely scenes where he may die.


So glad we had this discussion, as this is a vital analysis that needs to be done… stay tuned.