![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
Entry tags:
Come on Netflix, work with me here
So I just watched Monsters on Netflix streaming. A quite strong indie SF film that is a little too indie and not enough sci-fi for its own premise. Clearly wishes it was District 9/Cloverfield, but overall decent effort, especially for a film whose entire crew fit into a single van and whose director did all the special effects himself with "off-brand" Adobe Premier.
I watched the film because Netflix thought I would like it by 3.5/5 stars. They were absotively spot-on; this film fell exactly into the grey between the decent-and-enjoyable-but-nothing-special 3 and the excellent-not-quite-perfect-but-eminently-recommendable 4.
But Netflix doesn't let you rate things by half-stars. That is a privilege that they reserve purely for themselves.
Who knows why, exactly, they do this? But here's the problem:
Netflix prides itself on having a rating system that "learns" what an individual actually likes and adjusts its recommendations accordingly. So instead of proving a total aggregate rating of what all their users rated a particular film, Netflix provides an estimated rating of what you would likely rate the film, based on your usage and ratings history.
Now, I don't know the actual math involved beyond a reasonable assumption that it involves matrices, so I can't say for certain that being unable to use half-stars has a measurable effect or not. I would think it does, but I don't know for sure.
However, as a user of Netflix, I'm now frustrated because I don't know what I should rate this film. Should I rate it a 4, risking Netflix assuming it's really a 4.5 and recommending me the next Skyline as a safe 3? Should I rate it a 3, risking Netflix assuming a 2.5 and deciding to not recommend the next Moon because I'm clearly not too fond of indie SF?
When the interface you provide undermines how information is communicated, you have a useability problem, and it's past time Netflix addressed it.
I watched the film because Netflix thought I would like it by 3.5/5 stars. They were absotively spot-on; this film fell exactly into the grey between the decent-and-enjoyable-but-nothing-special 3 and the excellent-not-quite-perfect-but-eminently-recommendable 4.
But Netflix doesn't let you rate things by half-stars. That is a privilege that they reserve purely for themselves.
Who knows why, exactly, they do this? But here's the problem:
Netflix prides itself on having a rating system that "learns" what an individual actually likes and adjusts its recommendations accordingly. So instead of proving a total aggregate rating of what all their users rated a particular film, Netflix provides an estimated rating of what you would likely rate the film, based on your usage and ratings history.
Now, I don't know the actual math involved beyond a reasonable assumption that it involves matrices, so I can't say for certain that being unable to use half-stars has a measurable effect or not. I would think it does, but I don't know for sure.
However, as a user of Netflix, I'm now frustrated because I don't know what I should rate this film. Should I rate it a 4, risking Netflix assuming it's really a 4.5 and recommending me the next Skyline as a safe 3? Should I rate it a 3, risking Netflix assuming a 2.5 and deciding to not recommend the next Moon because I'm clearly not too fond of indie SF?
When the interface you provide undermines how information is communicated, you have a useability problem, and it's past time Netflix addressed it.
no subject