News for hackers. Stuff that matters.
Re:Netflix Replacing Star Ratings With Thumbs Ups and Thumbs... by subroutine on Friday March 17, 2017 @10:30AM (#13894217)
The goal isn't to broadly determine which movies are good and bad; the goal is to make accurate predictions about which titles a particular person will enjoy.

One good way to accomplish this is to pool users into cohorts of similar tastes, then suggest movies enjoyed by their cohort but not yet watched by a given user.

The question is whether doubling the frequency of rating a title is more valuable than a slightly more nuanced understanding of what each person likes. 'More valuable' meaning they can build you a better cohort from which to make title suggestions.

Re:Netflix Replacing Star Ratings With Thumbs Ups and Thumbs... by kartickv on Friday March 17, 2017 @11:54AM (#13894930)
Everything you've said makes sense, but I don't see it being contradictory to what I said.

If you decide you want a more nuanced understanding of what a given reviewer likes, I think it's better to evaluate different dimensions (Was it funny? Was it violent? etc) than shades of grey* of the same one.

* pun not intended :)