4 ms·

A Very Murray Christmas shows up as 3.5 stars for me. Note that the stars you see are how much Netflix thinks that YOU will like the show, not how much EVERYONE

A Very Murray Christmas shows up as 3.5 stars for me. Note that the stars you see are how much Netflix thinks that YOU will like the show, not how much EVERYONE likes the show.

dimillian 10y ago
And this is the most confusing part of Netflix for me. I'm very selective in what I watch, and Netflix partly know it, but the star system is wrong. I have to check IMDB every time when I see a show or a movie with an interesting title/plot but 1 or 2 stars. It's almost as if they don't want me to watch their content. I would prefer a lot better if it's displayed the average rating on Netflix or from IMDB for example.
- RangerScience 10y ago
  "The problem isn't thinking machines, it's letting machines do the thinking for you." Whether powered by algos or people, any rating system is "letting a machine do the thinking for you". Take the ten(?) minutes you spend researching the thing to watch ten minutes of the thing, and I bet you more often than not you'll know whether you want to watch it.
SyneRyder 10y ago
That makes sense. But in that case, it's bizarre to put 1-star recommendations in the "Top Picks For You" section. Feels like that section should be exclusively 4 & 5 star picks, no? [I'm glad someone got a higher rating for Very Murray Christmas, I like some of his movies. Perhaps I'll give it a try.]
- ghaff 10y ago
  It's "OK." Not remotely one of Murray's best films but hardly a waste of time. On the broader topic, I'm not sure anyone has really figured out how to do recommendations well whether film, music, TV, restaurants, or other products. And the more "clever" and personalized a system tries to be, the more likely it is to throw out some really bizarre recommendations.
  - SyneRyder 10y ago
    For movies, you could try MovieLens. I started using it a lot once Netflix's recommendations were becoming inaccurate for me. You can choose between 4 different algorithms, there's tagging to classify what you like/don't like about a movie, and lots of stats to help you understand the algorithms & your own taste. It's run by the University of Minnesota. https://movielens.org/home https://movielens.org/home
    - ghaff 10y ago
      Interesting. I'll give it a try. Someone I know did a lot of work on music recommendation and gave some presentations on it (including at SXSW as I recall). It's a harder problem than one might naively expect.

A Very Murray Christmas shows up as 3.5 stars for me. Note that the stars you see are how much Netflix thinks that YOU will like the show, not how much EVERYONE likes the show.

dimillian 10y ago

And this is the most confusing part of Netflix for me. I'm very selective in what I watch, and Netflix partly know it, but the star system is wrong. I have to check IMDB every time when I see a show or a movie with an interesting title/plot but 1 or 2 stars. It's almost as if they don't want me to watch their content. I would prefer a lot better if it's displayed the average rating on Netflix or from IMDB for example.

RangerScience 10y ago
"The problem isn't thinking machines, it's letting machines do the thinking for you." Whether powered by algos or people, any rating system is "letting a machine do the thinking for you". Take the ten(?) minutes you spend researching the thing to watch ten minutes of the thing, and I bet you more often than not you'll know whether you want to watch it.

SyneRyder 10y ago

That makes sense. But in that case, it's bizarre to put 1-star recommendations in the "Top Picks For You" section. Feels like that section should be exclusively 4 & 5 star picks, no? [I'm glad someone got a higher rating for Very Murray Christmas, I like some of his movies. Perhaps I'll give it a try.]

ghaff 10y ago
It's "OK." Not remotely one of Murray's best films but hardly a waste of time. On the broader topic, I'm not sure anyone has really figured out how to do recommendations well whether film, music, TV, restaurants, or other products. And the more "clever" and personalized a system tries to be, the more likely it is to throw out some really bizarre recommendations.
- SyneRyder 10y ago
  For movies, you could try MovieLens. I started using it a lot once Netflix's recommendations were becoming inaccurate for me. You can choose between 4 different algorithms, there's tagging to classify what you like/don't like about a movie, and lots of stats to help you understand the algorithms & your own taste. It's run by the University of Minnesota. https://movielens.org/home https://movielens.org/home
  - ghaff 10y ago
    Interesting. I'll give it a try. Someone I know did a lot of work on music recommendation and gave some presentations on it (including at SXSW as I recall). It's a harder problem than one might naively expect.