In another iteration of "equations will ruin art", this article talks about a successful statistical script doctoring process; [link]
Interestingly, (or depressingly) they cite two movies that appealed to me not in the least--Lincoln Lawyer would probably catch my eye on cable, but if it ended up on my Netflix queue, I wouldn't bump it to the top. And Oz just...I like the source material, but how more boring could it look?
So, while it makes sense that you can predict to a point box office and how to pander (it's not like that's a new concept--it's the spreadsheet in the background that appals), what does that to for me? If they could tell me what movies I liked they applied this to, and what changes were at their behest, I'd be fascinated.
But saying guardian superheroes do better than cursed one seems to, if I'm on the same semantic page, predict that Spiderman and Batman should do less well Superman (they cite him specifically), and the Avengers as a whole as successful as it was.
TL/DR: Send me contextualised relevant data, pls.